PIC.Robotics.A.Beginners.Guide.to.Robotics.Projects

PIC RoboticsThis page intentionally left blank.PIC RoboticsA Beginner’s Guide to Robotics Projects Using the PICmicro John Iovine McGrawHill New York Chicago San Francisco Lisbon London Madrid Mexico City Milan New Delhi San Juan Seoul Singapore Sydney TorontoCopyright © 2004 by The McGrawHill Companies, Inc. All rights reserved. Manufactured in the United States of America. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written permission of the publisher . 0071394559 The material in this eBook also appears in the print version of this title: 0071373241. All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every occurrence of a trademarked name, we use names in an editorial fashion only, and to the benefit of the trademark owner , with no intention of infringement of the trademark. Where such designations appear in this book, they have been printed with initial caps. McGrawHill eBooks are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate training programs. For more information, please contact George Hoare, Special Sales, at george_hoare@mcgrawhill.com or (212) 9044069. TERMS OF USE This is a copyrighted work and The McGrawHill Companies, Inc. (“McGrawHill”) and its licensors reserve all rights in and to the work. Use of this work is subject to these terms. Except as permitted under the Copyright Act of 1976 and the right to store and retrieve one copy of the work, you may not decompile, disassemble, reverse engineer, reproduce, modify, create derivative works based upon, transmit, distribute, disseminate, sell, publish or sublicense the work or any part of it without McGrawHill’s prior consent. You may use the work for your own noncommercial and personal use; any other use of the work is strictly prohibited. Your right to use the work may be terminated if you fail to comply with these terms. THE WORK IS PROVIDED “AS IS.” McGRAWHILL AND ITS LICENSORS MAKE NO GUARANTEES OR W ARRANTIES AS TO THE ACCURACY, ADEQUACY OR COMPLETENESS OF OR RESUL TS TO BE OBT AINED FROM USING THE WORK, INCLUDING ANY INFORMATION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHER WISE, AND EXPRESSLY DISCLAIM ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. McGrawHill and its licensors do not warrant or guarantee that the functions contained in the work will meet your requirements or that its operation will be uninterrupted or error free. Neither McGrawHill nor its licensors shall be liable to you or anyone else for any inaccuracy, error or omission, regardless of cause, in the work or for any damages resulting therefrom. McGrawHill has no responsibility for the content of any information accessed through the work. Under no circumstances shall McGrawHill and/or its licensors be liable for any indirect, incidental, special, punitive, consequential or similar damages that result from the use of or inability to use the work, even if any of them has been advised of the possibility of such damages. This limitation of liability shall apply to any claim or cause whatsoever whether such claim or cause arises in contract, tort or otherwise. DOI: 10.1036/0071394559For more information about this title, click here. Contents Preface xi Chapter 1. Robot Intelligence 1What Is a Microcontroller? 1Why Use a Microcontroller? 1Designer Computers—So Many Microcontrollers 2The Compiler 2PIC Programming Overview 2Software and Hardware 2PicBasic and PicBasic Pro Compilers 3EPIC Programmer 4Firmware 5Consumables 516F84 PIC Microcontroller 5Step 1: Writing Code (the Basic Program) 6Step 2: Using the Compiler 7Step 3: Installing the Firmware, or Programming the PIC Chip 7Ready, Steady, Go 8Parts List 8Chapter 2. Installing the Compiler 11Installing the PicBasic Compiler Software 11Installing the PicBasic Pro Compiler 12Chapter 3. Installing the EPIC Software 19Installing the EPIC Software in Windows 19Installing the EPIC Software, DOS Version 22Supplemental—Applications Directory 22Chapter 4. CodeDesigner 23CodeDesigner Features 23Software Installation 25vvi Contents Setting CodeDesigner Options 25First Program 31The EPIC Programming Board Software 33Parts List 35Chapter 5. Using DOS to Code, Compile, and Program 37Compile 39Programming the PIC Chip 43The EPIC Programming Board Software 44Using EPIC DOS Version 45Continuing with the wink.bas Program 45Chapter 6. Testing the PIC Microcontroller 49The PIC Microntroller 49The Solderless Breadboard 49Three Schematics, One Circuit 50Wink 53Troubleshooting the Circuit 54PIC Experimenter’s Board and LCD Display 54PIC Experimenter’s Board 56Use 56Simple Experiment 58Builtin LCD 60Using the LCD: PicBasic and PicBasic Pro Examples 63Introduction to Binary and the PIC Microcontroller 65Using TRIS and Port Registers 68Accessing the Ports for Output 70Counting Program 71Counting in Binary by 1 71Input 72The button Command 72A button Example 74peek 75peek and PicBasic Pro 76Basic Input and Output Commands 77Servomotors 78Parts List 79Chapter 7. Intelligence 83Approaches to Building Intelligence 83Where’s the Intelligence? 85Layered Behavioral Responses 85BehaviorBased Robotics 86Chapter 8. Walter’s Turtle 87BehaviorBased Robotics 87William Grey Walter—Robotics Pioneer 87Four Modes of Operation 88Contents vii Observed Behavior88 89 90 91 97 99 100 101 104 107 109 112 114 115 121 121 121 121 121 123 123 Building a Walter TortoiseDrive and Steering MotorsModifying the HS425BB ServomotorSheet Metal FabricationShellFinding the Center of GravityAttaching Bumper to Robot BaseBumper SwitchMounting the Steering ServomotorPhotoresistorTrimming the Sensor ArraySchematicProgramAdding Sleep ModePowerBehaviorFudge FactorLight IntensityHandednessParts ListChapter 9. Braitenberg Vehicles125 Neural I/O Relationships126 126 128 129 131 133 137 139 139 VehiclesBuilding VehiclesBack WheelsFront WheelsCdS Photoresistor CellsTrimming the Sensor ArrayPIC 16F84 MicrocontrollerTestingSecond Braitenberg Vehicle (Avoidance Behavior) 141Parts List 141Chapter 10. Hexapod Walker 143Imitation of Life 143Six Legs—Tripod Gait 143ThreeServomotor Walker Robot 144Function 144Moving Forward 145Moving Backward 146Turning Left 147Turning Right 148Construction 148Mounting the Servomotors 151Leg Positioning 152Linkage 154viii Contents Center (Tilt) Servomotor 154Sensors 155Electronics 158Microcontroller Program 159Parts List 164Chapter 11. Speech Recognition 165Applications 167Software Approach 167Learning to Listen 167SpeakerDependent and SpeakerIndependent Recognition 167Recognition Style 168Speech Recognition Circuit 168Circuit Construction 168Keypad 169To Train 169Testing Recognition 172Error Codes 172Clearing the Trained Word Memory 172Independent Recognition System 172Voice Security System 173Speech Interface Control Circuit 173How the Circuit Works 174PIC 16F84 Microcontroller Program 176Active High Output 177SPDT Relay Output 177Circuit Construction 177Programming the Speech Recognition Circuit: Training, Testing, and Retraining 177SRI02 and SRI03 Interfaces 180Robot Control 180Parts List 181Chapter 12. Robotic Arm 185Servomotor Building Blocks for Robotics 185Basic Servomotor Bracket Assembly 186Assembling MultipleServomotor Assemblies 189Building a FiveServomotor Robotic Arm 192Servomotors 197Servomotor Controllers 199Simple Servomotor Controller 200Fourand FiveServomotor Controllers 204Increasing the Lifting Capacity of the Robotic Arm 215Adding a Robotic Arm Base 216Parts List 223Chapter 13. Bipedal Walker Robot 225A Question of Balance? 226A Little Feedback 227Servomotors 227Contents ix Servomotor Brackets 228Footpads 228Assembly 230Schematic 231Program 233Subroutines M1, M2, and M3 234Going Further 241Turning Right and Left 242Parts List 242Chapter 14. Color Robotic Vision System 243CMU Camera 244Serial Communication 245VB Application Program 248Interfacing the CMU Camera to a Robot 250PIC 16F84 Runs at 16 MHz 251Program 1 251Program 2 254Incandescent or Fluorescent Lighting 254Servomotors for Robot 259Program 3 261Robot Construction 264Running the Program 266Going Further 267Parts List 267Suppliers 269Index 271This page intentionally left blank.PrefaceThis is a project book on building small robots. Each robot utilizes the PICmicro series of microcontrollers from Microchip Technologies Inc. for intelligence, navigation, motor control, and sensory readings. By changing the microcontroller programming and sensory electronics we can create a zoo of robots that includes photovores, behaviorbased (neural) robots, hexapod and bipedal walkers, and artificial vision systems that can track and follow objects. Each robot project has something to teach. John Iovine xiThis page intentionally left blank.1 Chapter Robot Intelligence The robotic projects outlined in this book make extensive use of the PIC series of microcontroller from Microchip T echnology Inc. In addition to its ability to run programs, the microcontroller has input and output lines (pins) that are used to control motor drive systems, read sensors, and communicate. W e demand a lot from our microcontroller(s), so it’s important to have a good idea of what a microcontroller is right from the start. What Is a Microcontroller? A microcontroller is essentially an inexpensive singlechip computer . Single chip means the entire computer system lies within the confines of a sliver of silicon encapsulated inside the plastic housing of an integrated circuit. The microcontroller has features similar to those of a standard personal computer. The microcontroller contains a CPU (central processing unit), RAM (random access memory), ROM (readonly memory), I/O (input/output) lines, serial and parallel ports, timers, and sometimes other builtin peripherals such as ana logtodigital (A/D) and digitaltoanalog (D/A) converters. The key feature, however, is the microcontroller’s capability of uploading, storing, and running a program. Why Use a Microcontroller? Being inexpensive singlechip computers, microcontrollers are easy to embed into larger electronic circuit designs. Their ability to store and run unique programs makes them extremely versatile. For instance, one can program a microcontroller to make decisions and perform functions based on situations (I/O line logic) and events. The math and logic functions allow the microcontroller to mimic sophisticated logic and electronic circuits. Copyright © 2004 The McGrawHill Companies. Click here for terms of use. 12 Chapter One Programs can also make the microcontroller behave as a neural network and/or a fuzzy logic controller . Microcontrollers are incorporated in consumer electronics and are responsible for the “intelligence” in these smart electronic devices. Designer Computers—So Many Microcontrollers There are a large variety of microcontrollers on the market. W e will use the versatile microcontroller chips called PIC chips (or PICmicro chips) from Microchip Technology Inc. The CompilerThere are a number of compilers on the market that allow users to write programs (code) in different highlevel languages. Highlevel language frees the programmer from wrestling with and controlling the microcontroller ’s registers when writing code and accessing the different aspects of the microcontroller’s features and memory. The highlevel language I use is a derivative of the Basic language. It is called PicBasic. (The PicBasic and PicBasic Pro compilers used to write PicBasic programs are products and trademarks of microEngineering Labs, Inc.) PicBasic is similar to the PBasic language used in programming the Basic Stamp series. Programming microcontrollers directly using the PicBasic (or PicBasic Pro) compiler offer two major advantages over the Basic Stamp series of microcontrollers which use external serial EEPROM for memory storage, faster program execution speed (20to 100fold increase), and reduced cost. PIC Programming Overview Programming PIC microcontrollers is a simple threestep process: W rite the code, compile the code, and upload the code into a microcontroller . Following is an overview of the process; stepbystep instructions will be provided in the following chapters. Software and Hardware You will need two items to begin programming and building microcontroller based projects and robotics. First is the compiler , either the PicBasic Pro or PicBasic compiler (see Fig. 1.1). The PicBasic Pro compiler from microEngineering Labs, Inc. has a suggested retail price of $249.95. The PicBasic compiler from microEngineering Labs, Inc. has a suggested retail price of $99.95. In addition to a compiler you need the EPIC programming board and software; this package sells for $59.95 (see Fig. 1.2). (EPIC is a product and trademark of microEngineering Labs, Inc.)PicBasic and PicBasic Pro Compilers The PicBasic and PicBasic Pro compilers both function in the same way. Saved program code (text file) is run through a compiler (either the PicBasic or PicBasic Pro compiler). The compiler reads through the text file and creates (compiles) an equivalent machine code instruction listing ( .hex file) of the program. The machine code (.hex file) is a list of hexadecimal numbers that represent the PicBasic program. The list of hexadecimal numbers ( .hex file) is uploaded (programmed) into the microcontroller . When the microcontroller Robot Intelligence 3 Figure 1.1 PicBasic Pro and PicBasic software packages and manuals. Figure 1.2 EPIC Programmer software and hardware. 4 Chapter Oneis started, its CPU will run through the programmed list of hexadecimal numbers, running the PicBasic program. Uploading the machine code ( .hex file) into the microcontroller is the job of the EPIC Programmer board and software, which we will look at shortly. The PicBasic Pro compiler is considerably more expensive than the standard PicBasic compiler. The Pro version offers an enhanced and richer basic command syntax than is available in the PicBasic compiler package. A few of the additional commands that can be found in the Pro version allow the use of interrupts, direct control of LCD modules, DTMF out, and X10 commands, to name a few . While the PicBasic Pro is a more sophisticated package, the compiler does not handle two of my favorite Basic commands, peek and poke. Although the commands are listed as “functional” in the Pro manual, it is emphasized that “PEEK and POKE should never be used in a PicBasic Pro program.” There are workarounds to using the peek and poke commands in the Pro version that will be covered when needed later on. In the balance of this book, at times I will refer to both the PicBasic and PicBasic Pro compilers simply as the compiler(s). This saves me from continually writing PicBasic and PicBasic Pro compiler throughout the book. When a distinction becomes necessary, I will specify the individual compiler. The compiler program may be run manually in DOS or in an “MSDOS Prompt” window. A third option, and one you will probably use, is to run the compiler within a W indows program called CodeDesigner . CodeDesigner is discussed later in this chapter and fully in Chap. 4. The minimum system requirement for the compiler is an XT class personal computer (PC) running DOS 3.3 or higher. The compiler can compile programs for a large variety of PIC microcontrollers. EPIC Programmer The second item needed is the EPIC Programmer , also made by microEngineering Labs, Inc. The EPIC Programmer consists of software (EPIC) and a programming carrier board (hardware). The EPIC software package has two executable files, one for DOS and another version of the software for Windows. It is the EPIC hardware and software that takes the compiled .hex file generated by the compiler and uploads it into the microcontroller, where it may be run. The EPIC Programmer is compatible with both the PicBasic and PicBasic Pro compilers. The programming carrier board (see Fig. 1.3) has a socket for inserting the PIC chip and connecting it to the computer , via the printer port, for programming. The programming board connects to the computer ’s printer port via a DB25 cable. If the computer only has one printer port with a printer connected to it, the printer must be temporarily disconnected to program PIC chips. The EPIC programming carrier board supports a large variety of PIC microcontrollers.Firmware Many writers use the term firmware. This word is used when software is embedded in a hardware device that can read and execute by the device but cannot be modified. So when our program (software) is embedded (uploaded) into the microcontroller, it may be referred to as firmware. Other phrases may include the term firmware instead of software, such as “upload the firmware” or “once the firmware has been installed into the device.” ConsumablesConsumables are the electronic components, the PIC microcontroller chip itself, with a few support components to get the microcontroller up and run ning. I recommend beginning with the 16F84 PIC microcontroller . The 16F84 is an 18pin dip chip with 13 I/O lines and has 1K 3 14 of rewritable memory. The rewritable memory allows you to reprogram the PIC chip up to 1000 times to test and troubleshoot your programs and circuits. The minimal support components are a 5V dc power supply , oscillator (4.0MHz crystal), and one pullup 1/4W resistor (4.7kV). 16F84 PIC Microcontroller The PIC 16F84 microcontroller is shown in Fig. 1.4. It is a versatile microcontroller with flash memory. Flash memory is the terminology used to describe “rewriteable” memory. The 1K 3 14bit onboard flash memory can endure a Robot Intelligence 5 Figure 1.3 Closeup of EPIC programming carrier board. 6 Chapter OneFigure 1.4 Pinout of 16F84 PIC microcontroller integrated circuit. General features: RISC CPU 35 singleword instructions; operating speed dc, 10MHz clock input; 1K program memory; 14bitwide instructions; 8bitwide data path; direct, indirect, and relative addressing; 1000 erase/write cycles. Peripheral features: 13 I/O pins with individual direction control; highcurrentsink/source for direct LED drive (25mA sink max. per pin, 20mA source max. per pin); TMRO—8bit timer/counter with 8bitprogrammable prescaler.minimum of 1000 erase/write cycles. So you can reprogram and reuse the PIC chip at least 1000 times. The program retention time between erase/write cycles is approximately 40 years. The 18pin chip devotes 13 pins to I/O. Each pin may be programmed individually for input or output. The pin status (I/O direction control) may be changed on the fly via programming. Other features include power on reset, power saving sleep mode, power up timer, and code protection. Additional features and architectural details of the PIC 16F84 will be given as we continue. Step 1: Writing Code (the Basic Program) Both the PicBasic and PicBasic Pro compilers are packaged with a free version of CodeDesigner software. CodeDesigner is an integrated development environment (IDE) for writing and programming PIC microcontrollers. CodeDesigner is an advanced text editor that is capable of calling and using both the PicBasic and PicBasic Pro compilers and the EPIC software.If you don’t want to use CodeDesigner, program text files may be written using any word processor as long as it is able to save its text file as ASCII or DOS text. If you don’ t own a commercial word processor , you can use Windows Notepad, which is included with W indows 3.X, 95, and 98. If you work at the DOS level, you can use the Edit program to write text files. When you save the text file, save it with a .bas suffix. For example, if you were saving a program named wink, save it as wink.bas. Step 2: Using the Compiler Once set up, the CodeDesigner software will call and control the compiler and programmer software. The compiler may be run manually from a DOS window. To run the compiler program manually, enter the command pbc followed by the number of the PIC chip being programmed (that is, 16F84), then followed by the name of the source code text file. For the PicBasic Pro compiler program, the command starts with pbp instead of pbc, followed by the name of the source code text file. For example, for the PicBasic compiler, if the source code text file we created is named wink, then at the DOS command prompt enter pbc p16f84 wink.bas For the PicBasic Pro compiler, the command line would be pbp p16f84 wink.bas The compiler reads the text file and compiles two additional files, an .asm (assembly language) and a .hex (hexadecimal) file. The wink.asm file is the assembly language equivalent to the Basic program. The wink.hex file is the machine code of the program written in hexadecimal numbers. It is the .hex file that is uploaded into the PIC chip. If the compiler encounters errors when compiling the PicBasic source code, it will issue a list of errors it has found and will terminate. The errors listed need to be corrected in the source code (text file) before it will successfully compile. Step 3: Installing the Firmware, or Programming the PIC Chip Connect the EPIC programming board to the computer ’s printer port via a DB25 cable. If you are using CodeDesigner , launch the EPIC Programmer from the menu. The EPIC programming board must be connected to the parallel port and switched on before you start the software, or else the software will issue an error message “EPIC Programmer not found.” Aside from the EPIC Windows software (epicwin.exe), which may be started manually in Windows or through the CodeDesigner software, there is also a DOS version of the program called epic.exe. Robot Intelligence 7 8 Chapter OneFigure 1.5 Windows version of EPIC software. Figure 1.5 is a picture of the EPIC W indows program screen. Use the Open File option and select wink.hex from the files displayed in the dialog box. The file will load and numbers will be displayed in the code window on the left. Insert the 16F84 into the socket on the programming board, and select the Program option from the Run menu. An alternative to using the menu option is to press the Ctrl and P buttons on the keyboard. The software is then uploaded into the PIC microcontroller and is ready to be inserted into your circuit and go to work. Ready, Steady, Go Subsequent chapters contain stepbystep instructions for installing the software onto your hard drive and programming your first PICmicro chip. Parts List PicBasic Pro compiler $249.95 PicBasic compiler 99.95 EPIC Programmer 59.95 Microcontroller (16F84) 7.95Robot Intelligence 9 6ft cable (DB25) 6.95 4.0MHz Xtal 2.50 (2) 22pF capacitors 0.10 each Available from Images SI Inc. (see Suppliers at end of book). Additional components are required in Chap. 6: (1) Solderless breadboard RadioShack PN# 276175 (1) 0.1mF capacitor RadioShack PN# 2721069 (8) Red LEDs RadioShack PN# 276208 (8) 470V resistors* RadioShack PN# 2701115 (1) 4.7kV resistor RadioShack PN# 2711126 (1) Voltage regulator (7805) RadioShack PN# 2761770 (2) Fourposition PC mounted switches RadioShack PN# 2751301 (1) 9V battery clip RadioShack PN# 270325 Available from RadioShack, Images SI Inc., Jameco Electronics, and JDR Microdevices (see Suppliers). *These resistors are also available in 16pin dip package.This page intentionally left blank. 2 Chapter Installing the Compiler To compile your PicBasic programs (text files) into something that can be uploaded into the PIC microcontrollers and run, you need to first run the program text file through a compiler. So the first step is to load the compiler software onto your computer’ s hard drive . The following are instructions for installing the PicBasic compiler. A section on installing the PicBasic Pro com piler follows these instructions. Installing the PicBasic Compiler Software The first thing you need to do is to create a subdirectory on your computer’ s hard drive for the PicBasic compiler softw are. I will use Windows Explorer (Windows 95, 98, ME, 2000, XP) to create this directory.Windows Explorer can be found in the Programs folder in Windows 95 and 98 (see F ig. 2.1). For Windows ME, 2000, and XP users , Windows Explorer can be found in the Accessories folder (see Fig. 2.2). Create a subdirectory called PBC on the computer’ s hard drive; then copy the files from the diskette into it. For the conventions in this book it is assumed that the reader’s hard drive is drive letter C . Start the Windows Explorer program. Highlight your computer’s hard drive (usually the C drive) in the Folders window. Next highlight the File menu, then New menu, and click on the Folder option (see Fig. 2.3). Enter the name PBC in the New Folder icon (see Fig. 2.4). Place the 3.5in PicBasic compiler diskette into your computer’s floppy drive, usually the A drive. Highlight the A drive in Windows Explorer’s Folder window (see F ig. 2.5). All the files on the 3.5in diskette will be displa yed in the rightside area. Select all the files , go to Edit menu options , and choose Copy (see Fig. 2.6). Next select the PBC directory on the left side of the Windows Explorer window. Then go back to the Edit menu and select the P aste option. All the files and subdirectories on the 3.5in diskette will be copied into the PBC directory on the hard drive . Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 1112 Chapter Two Figure 2.1 Finding Windows Explorer in Windows 95 and 98. An alternate to pasting the selected files is to select all the files as before , copy the files, drag the selected files to the PBC directory using the mouse, and then release the mouse button (see F ig. 2.7). Installing the PicBasic Pro Compiler Installing the PicBasic Pro compiler is not the same procedure as outlined for the PicBasic compiler. To install the PicBasic Pro compiler, you must execute a selfextracting program that decompresses the necessary programs and files . It is recommended that you create a subdirectory named PBP on your computer’s hard drive. Start the Windows Explorer program. Highlight your computer’s hard drive (usually the C drive) in the Folders window. Next highlight the File menu, then New menu, and click on the Folder option (see F ig. 2.3). Enter the name PBP in the New Folder icon (see Fig. 2.4). Place the 3.5in PicBasic Pro Compiler diskette into your computer’ s floppy drive, usually the A drive . Now here’ s where the installation procedure changes. For those using Windows 95 or 98, start an MSDOS Prompt window. Click on Start, select Programs, then click on MSDOS Prompt (see F ig. 2.8). For Windows ME, 2000, and XP users , start a Command Prompt windowFigure 2.2 Figure 2.3 Finding Windows Explorer in Windows 2000 and XP. Creating a new folder (subdirectory) on computer’s hard drive C.Figure 2.4 Type subdirectory’s name in the New Folder icon. Figure 2.5 Selecting the A drive containing the PicBasic program diskette . 14Figure 2.6 Selecting and copying all files and subdirectories on the PicBasic program diskette. Figure 2.7 Using mouse to copy all selected files on the PicBasic program diskette in the A drive to the PBC directory on the hard drive. 1516 Chapter Two Figure 2.8 Starting MSDOS Prompt window in Windows 95 and 98. (equivalent to an MSDOS Prompt window). Click on All Programs , select Accessories, and then click on Command Prompt (see Fig. 2.9). In either the Command Prompt window or MSDOS window , you will need to type in and use a few oldfashioned DOS commands . DOS commands are typed in on the command line , and then the Return key is hit to execute the command. The DOS instructions are provided to help the reader and serve as a supplement to the installation directions provided with the softw are pac kages. The instructions are not meant as a DOS tutorial. More information on DOS commands can be found in any number of DOS manuals. Here is a list of DOS commands we will be using and what action they perform: Command Action cd Change directory md Make directory copy Copy files xcopy Copy files and subdirectories path Set a search path for executable files dir DirectoryFrom this point on, the MSDOS Prompt window and the Command Prompt window will be referred to as the DOS window . When the DOS window is opened, you will be located in a subdirectory on the hard drive . Your prompt may look like this: C:\WINDOWS.. The DOS prompt provides vital information. The C: tells us we are on the C drive. The \WINDOWS tells us we are in the Windows subdirectory. We want to work from the root directory of the computer’ s hard drive (usually the C drive).We accomplish this by using the cd (change directory) command. The cd.. command brings one up a single level in the directory hierarc hy. Using the cd\ command brings one up to the root directory regardless of how deep (levels) one has moved into subdirectories. The root directory is the top of the directory hierarc hy. From the Windows subdirectory type in cd\ and hit the Enter key to move to the root directory of the hard drive . Type in (enter) the following command and hit the Enter key. cd\ We already created our subdirectory PBP by using Windows Explorer for the PicBasic Pro compiler. We want to move into the PBP subdirectory , enter the following command, and hit Enter. Installing the Compiler 17 Figure 2.9 Starting Command Prompt window in Windows 2000 and XP.18 Chapter TwoFigure 2.10 Using DOS commands in DOS Prompt window to execute PicBasic Pro installation program.c:\> cd pbpNext place the 3.5in PicBasic Pro diskette into your A drive, and type the following at the DOS prompt: c:\pcp> a:\pbpxxx dHere xxx is the version number of the compiler on the disk (see Fig. 2.10). This command copies and installs all the required files into the PBP directory.With the files safely loaded onto your hard drive, remove the diskette and store it in a safe place, in case it is needed in the future . The PicBasic Pro program is now installed. You may close the DOS window and store the 3.5in diskette.3 Chapter Installing the EPIC Software Installing the EPIC softw are from Windows is simple . To install, run the install.bat file on the 3.5in EPIC diskette . The install.bat file exe cutes the main selfextracting program that automatically creates a subdirec tory EPIC on your computer’s hard drive, then decompresses the program and its support files, and copies them into the EPIC subdirectory. If a subdirectory called EPIC already exists on your hard drive , when you run the install.bat file, you will receive an error message . If you are still in the same DOS session as in last c hapter and want to continue to use DOS to install the EPIC software, skip down to the “Installing the EPIC Software, DOS Version” section. For those who wish to use Windows to install this software, continue to read. Installing the EPIC Software in Windows From Windows c lick on the Start button, then Run (see F ig. 3.1). Place the EPIC programming diskette into the A drive. When the Run menu window opens, select Browse. From the Browse window select the A drive. This action will list the files on the A drive. Select the install.bat file and c lick the Open button (see Fig. 3.2). This action brings you bac k to the Run window . The install.bat file should be listed in the window. See Fig. 3.3. Click on the OK button.This action automatically opens a DOS window and starts the executable program on the EPIC diskette. The executable program creates a new subdirectory on the computer’s hard drive called EPIC . It decompresses and copies all the necessary files into the EPIC subdirectory, as shown in Fig. 3.4. If you just installed the EPIC program using Windows, skip over the next section, “Installing the EPIC Software, DOS Version” and continue reading at the “Supplemental—Applications Directory” section. Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 1920 Figure 3.1 Selecting Run option on Windows start menu. Figure 3.2 Selecting install.bat file on EPIC diskette.Figure 3.3 Hitting OK on install.bat to begin execution. Figure 3.4 Selfextracting EPIC program running in MSDOS window.22 Chapter Three Installing the EPIC Software, DOS Version If you are still operating in the same DOS session as in Chap . 2, move back into the root directory, and enter at the prompt c:\> pbp cd..If you are entering a new DOS window , the prompt may appear a little different, but the command is the same. c:\> windows cd/From the root directory of the C drive we will run the install.bat program on the EPIC 3.5in diskette . The selfextracting file creates itself a subdirectory called EPIC. Place the 3.5in EPIC diskette into the floppy drive . At the DOS prompt enter c:\> a:This places the command prompt into the A drive; the command prompt should look like this: a:\>Now run the install.bat file by entering the following command: a:\> install.batThis starts the selfextracting file that creates the EPIC subdirectory and installs all the required files into the subdirectory. With the program and files installed onto your hard drive , remove the diskette and store it in a safe place , in case it is needed in the future . Supplemental—Applications Directory It would be a good idea at this time if we created another subdirectory where we can store all our PicBasic application programs. This will keep the PBC (or PBP) and EPIC directories c lean, neat, and uncluttered with programs and program revisions. From Windows Explorer create an Applics subdirectory on your computer’ s hard drive.4 Chapter CodeDesigner In this c hapter we will set up and work with the CodeDesigner softw are. CodeDesigner is a Windows integrated development environment (IDE) interface for the PIC series of microcontrollers . This IDE interface allows one to write code, compile the code, and then program the code into a PIC microcontroller while staying in the same Windows environment. The compiling of code within CodeDesigner still requires the use of one of the PicBasic compilers. Programming the compiled code into a PIC microcontroller requires the EPIC softw are and hardw are. CodeDesigner integrates these softw are and hardw are pac kages so that they can operate within its Windows environment. CodeDesigner has many useful features that help you write code and that make it far superior to using a simple text editor . CodeDesigner Features ■ AutoCodeCompletion. CodeDesigner makes writing code muc h easier with smart popup list boxes that can automatically fill in statements and para meters for you. ■ Multidocument support. ■ Line error highlighting . Compile your PicBasic project and CodeDesigner will read error data and highlight error lines . ■ QuickSyntaxHelp. The QuickSyntaxHelp feature displays statement syntax when you type in a valid PicBasic statement. ■ Statement description. Statement descriptions are displa yed in the status bar when you type in a valid PicBasic statement. ■ Statement Help. Simply position your cursor over a PicBasic statement and get statementspecific help. Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 2324 Chapter Four ■ Label listbox. The label listbox displa ys the current label and allows you to select a label from the list to jump to . ■ Colored PicBasic syntax. This sets colors for reserved words , strings, numbers, comments, defines, etc. Colored PicBasic syntax makes for easy code reading. ■ Bookmarks. Never lose your place again. CodeDesigner allows you to set bookmarks. ■ Multiple undo/redo actions. If you didn’t want to delete that last line, it’s no problem. Simply click on the Undo button. ■ Multiple views. Multiple views of your source code allow you to easily edit your code. ■ Print source code. ■ Draganddrop text. ■ Row/columnbased Insert, Delete, and Copy. ■ Search and replace. ■ Compile and launch device programmer. One feature I like is that eac h typed line of code is color coded, making it easier to spot errors and read through your code . When you purchase either the PicBasic or PicBasic Pro compilers, it is packaged with an additional diskette that contains a free version of CodeDesigner called CodeDesigner Lite. The Lite version allows you to write programs up to 150 lines and open up three source files at once for easy copy and paste . If you would like to try CodeDesigner without purc hasing a compiler, CodeDesigner Lite is freely downloadable from the Internet (see Parts List at end of chapter). The idea is, if you like the free CodeDesigner software, you can then upgrade to the fullfeatured CodeDesigner. The standard version of CodeDesigner costs $75.00 and removes the restrictions imposed in the Lite version.This standard version allows you to write programs with an unlimited amount of code lines and to open an unlimited amount of source files . Of course, unlimited means with respect to the limits of your computer’ s capabilities. If for any reason someone does not wish to use the CodeDesigner softw are, the procedures for writing code and compiling and programming a PICmicro chip manually from a DOS environment are covered in Chap . 5. CodeDesigner increases productivity and the ease with which you can write, debug, and load PicBasic programs into the microcontroller. If there is a problem (more often than not), debugging the code and recompiling are muc h easier and faster when you are using CodeDesigner . When the program is completely debugged, it can be uploaded into the PIC microcontroller via the EPIC software and programming board. At this point the microcontroller and circuit are tested. If they function properly, I’m finished; if not, I begin rewriting the program or redesigning the electronics .CodeDesigner 25 Software Installation The CodeDesigner softw are loads as most standard Windows software does. Load the software on your computer’s hard drive according to the instructions given with the software. When CodeDesigner installs, it creates a subdirectory in the Program F iles directory. It places a CodeDesigner shortcut on the Start, Program menu in Windows. Setting CodeDesigner Options In order for CodeDesigner to compile code and program the resulting code into PIC microcontrollers , we need to configure the default locations where CodeDesigner looks for its support programs . We set up the default locations by entering the software paths where CodeDesigner stores programs, looks for the PicBasic compiler and where to find the EPIC program. Start the CodeDesigner software (see Fig. 4.1); the evaluation copy opens with this version of the window. The next window is the standard opening screen to the CodeDesigner softw are (see F ig. 4.2). To begin setting the options , click on the Compile menu option and then on Compiler Options (see F ig. 4.3). Figure 4.1 CodeDesigner Lite version startup screen.26 Chapter Four Figure 4.2 Program written in CodeDesigner ready for compiling and programming. Figure 4.3 Error created when Windows cannot identify EPIC Programmer attached to printer port.CodeDesigner 27 Figure 4.4 Schematic for wink program. The Compiler Options window opens (see F ig. 4.4). In the top text field, use the pulldown menu to select whic h compiler you are using , the PicBasic Pro or PicBasic. In Fig. 4.4 the PicBasic Pro compiler is c hosen. In the second text field, you select the compiler’ s pathname. The compiler path and name (pbpw.exe) is chosen for the PicBasic Pro compiler, in the subdirectory of C:\PBP. In the third text field we c hoose where the CodeDesigner softw are looks to load and sa ve our source code files . Hit the Browse button next to the text field. This opens a browser window (see F ig. 4.5); select the Applics subdirectory on the hard drive and c lick on OK. The “Default Source Code Directory” text field now contains the path “C:\Applics” subdirectory (see F ig. 4.6). Click the OK button to c lose the Compiler Options window . Now we need to set the Programmer Options . Click on the Programmer , Programmer Options on the top menu (see F ig. 4.7). This opens the Programmer Options window (see Fig. 4.8). Click on the Browse button next to the Programmer P athname text field. A browser window opens; select the epicwin.exe program in the EPIC subdirectory on your computer’s hard drive (see F ig. 4.9). Click Open, and this brings you bac k to the Programmer Options window . The new path you select should be in the Programmer Pathname text field (see Fig. 4.10). Click OK to set this option.Figure 4.5 Complete circuit built on solderless breadboard. Figure 4.6 Selecting compiler (CodeDesigner). 28Figure 4.7 Opening Programmer options (CodeDesigner). Figure 4.8 Programmer options window (CodeDesigner).Figure 4.9 Selecting epicwin.exe Programmer in EPIC subdirectory. Figure 4.10 Hitting OK to confirm selection. 30CodeDesigner 31 With the CodeDesigner options set, we are ready to write our first program. First ProgramStart CodeDesigner and enter the following code for the PicBasic compiler . ‘First PicBasic program to wink two LEDs connected to port b.loop: high 0 ‘Turn on LED connected to pin rb0 low 1 ‘Turn off LED connected to pin rb1 pause 500 ‘Delay for .5 second low 0 ‘Turn off LED connected to pin rb0 high 1 ‘Turn on LED connected to pin rb1 pause 500 ‘Delay for .5 second goto loop ‘Go back to loop and blink & wink LEDs forever end The next program is identical in function (not code) to the PicBasic program above. Start CodeDesigner and enter the following code (see F ig. 4.11): ‘Wink program‘Blinks and winks two LEDs connected to port bFigure 4.11 PicBasic Pro program written in CodeDesigner.32 Chapter Four loop: high portb.0 ‘Turn on LED connected to rb0 low portb.1 ‘Turn off LED connected to rb1 pause 500 ‘Wait 1/2 second low portb.0 ‘Turn off LED connected to rb0 high portb.1 ‘Turn on LED connected to rb1 pause 500 ‘Wait 1/2 second goto loop ‘Loop backrepeat cycle blink & wink forever CodeDesigner defaults to writing code for the PIC 16F84 microcontroller. This is the microcontroller I recommend to start with. To c hange the microcon troller, simply pull down the device menu and select the appropriate microcontroller (see Fig. 4.12). When CodeDesigner attempts to compile a program from the Windows environment, it automatically opens a DOS Prompt window, compiles the program, and then ends the DOS session. To compile the program using CodeDesigner, either select compile under the Compile menu or hit F5. CodeDesigner automatically starts the PicBasic Pro compiler (or PicBasic compiler) to compile the program. Before you attempt to compile a program, make sure you ha ve set up Compiler Options under the Compile menu. Figure 4.12 Pulldown menu location for selecting microcontroller.CodeDesigner 33 Figure 4.13 Error message generated when CodeDesigner cannot find Programmer. Once the program is compiled, we can go to the next step of loading the program into a PIC microcontroller c hip. Connect the EPIC programming board to the printer port. If your computer has only one printer port,disconnect the printer, if one is connected,and attach the EPIC programming board, using a 6ft DB25 cable. When you connect the EPIC programming board to the computer , there should not be any microcontroller installed onto the board. If you have an ac adapter for the EPIC programmer board, plug it into the board. If not, attach two fresh 9V batteries . Connect the “battery on” jumper to apply power. The programming board must be connected to the printer port with power applied to the programming board before the EPIC programming softw are is started. If it is not, the software will not see the programming board connected to the printer port and will give the error message “EPIC Programmer Not F ound” (see Fig. 4.13). The EPIC Programming Board Software To program the 16F84 microcontroller from within CodeDesigner , select the Launch Programmer menu item from the Programmer menu, or hit F6. CodeDesigner automatically starts the epicwin.exe Windows software.34 Chapter Four Figure 4.14 Setting configuration switches in epicwin.exe program. With the EPIC Windows software started, set the configuration switches one by one under the Options menu item (see F ig. 4.14). Device: Sets the device type. Set it for 16F84 (default). Memory Size (K): Sets memory size. Choose 1. OSC: Sets oscillator type. Choose XT for crystal. Watchdog Timer: Choose On. Code Protect: Choose Off. Power Up Timer Enable: Choose On. After the configuration switches are set, insert the PIC 16F84 microcontroller into the open socket on the EPIC programming board. If you received an error message “EPIC Programmer Not F ound” when CodeDesigner started the EPIC Windows program (see F ig. 4.13), you ha ve the option of either troubleshooting the problem or using the EPIC DOS program. For instructions on using the EPIC software, DOS version, see Chap. 5. The schematic of the circuit needed to test the PICmicro is given in Chap. 6. If you ha ve successfully written, compiled, and uploaded the code into the PICmicro c hip using CodeDesigner , then you can skip the DOS material in Chap. 5 and pick up at “Testing the PIC Microcontroller” in Chap. 6.CodeDesigner 35 Parts List CodeDesigner Lite Free Download from Internet at: www.imagesco.com/catalog/pic/codedesigner.html. CodeDesigner Standard $75.00 Available from Images SI Inc. (see Suppliers at end of book).This page intentionally left blank.5 Chapter Using DOS to Code, Compile, and Program In Chap . 4 we compiled and programmed our microcontroller , using the CodeDesigner program. If for some reason you do not wish to use or cannot use CodeDesigner Lite, this c hapter will instruct you in how to perform all the functions for writing code, compiling the code, and programming the code in a PICmicro chip from DOS or a DOS Prompt window. When you start a new DOS session, use the path command (see Fig. 5.1), so that you will not have to copy and swap files back and forth across directories. If you have created the directory names as suggested in this book, you can use the following command. For PicBasic users the command is path \;c:\pbc;c:\epic;c:\windows\command;For PicBasic Pro users, the command is path \;c:\pbp;c:\epic;c:\windows\command;Now we can begin by using a standard word processor or text editor to write the PicBasic program in DOS . Windows users can use the Notepad program. DOSlevel users can use the Edit program. In DOS we will work from and store our program(s) in the subdirectory we created earlier , called Applics. Move into the Applics subdirectory. Use the cd (change directory) command. Enter this at the DOS prompt (see F ig. 5.1): c:\> cd applicsOnce in this directory, the prompt changes to c:\applics>Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 3738 Chapter Five Figure 5.1 Entering DOS commands for path, changing directories, and starting the Edit program. In this example I will be using the free Edit program pac kage with Windows to write the program. Start Edit by typing Edit at the command prompt (see Fig. 5.1). c:\applics> editThis starts the Edit program (see F ig. 5.2). Enter this program in your word processor exactly as it is written: ‘1st PicBasic program‘Wink two LEDs connected to port b.loop: high 0 ‘Turn on LED connected to pin rb0 low 1 ‘Turn off LED connected to pin rb1 pause 500 ‘Delay for .5 second low 0 ‘Turn off LED connected to pin rb0 high 1 ‘Turn on LED connected to pin rb1 pause 500 ‘Delay for .5 second goto loop ‘Go back to loop and blink & wink LEDs forever end See Fig. 5.3. Save the text file in the Applics directory. Use the Save function under the File menu. Name the file wink.bas (see Fig. 5.4). If by accident you saved the file as wink.txt, don’t get discouraged. You can do aUsing DOS to Code, Compile, and Program 39 Figure 5.2 Start screen of Edit program. Save As from the Edit program (under F ile menu) and rename the file wink.bas.For PicBasic Pro users, enter the following text in your word processor and save the file as wink.bas. ‘1st PicBasic Pro program‘Winks two LEDs connected to port bloop:high portb.0 ‘Turn on LED connected to rb0low portb.1 ‘Turn off LED connected to rb1pause 500 ‘Wait 1/2 secondlow portb.0 ‘Turn off LED connected to rb0high portb.1 ‘Turn on LED connected to rb1pause 500 ‘Wait 1/2 secondgoto loop ‘Loop backrepeat cycle blink & wink foreverCompile The PicBasic compiler (or PicBasic Pro compiler) ma y be run from DOS or from a DOS Prompt window within Windows.40 Chapter Five Figure 5.3 Entering wink.bas program. Figure 5.4 Saving wink.bas program.Using DOS to Code, Compile, and Program 41 Figure 5.5 Entering DOS command to run PicBasic compiler program on the wink.bas program for the 16F84 microcontroller. We will run the PicBasic compiler from the Applics directory, type the command pbc p16f84 wink.bas at the DOS prompt, and hit the Enter key (see Fig. 5.5). c:\applics> pbc p16f84 wink.bas(For PicBasic Pro the command is c:/applics>pbp p16f84 wink.bas.) The compiler displays an initialization copyright message and begins processing the Basic source code (see F ig. 5.6). If the Basic source code is without errors (and why shouldn’t it be?), it will create two additional files. If the compiler finds any errors, a list of errors with their line numbers will be displayed. Use the line numbers in the error message to locate the line number(s) in the .bas text file where the error(s) occurred. The errors need to be corrected before the compiler can compile the source code correctly . The most common errors are with Basic language syntax and usage . You can look at the files by using the dir directory command. Type dir at the command prompt c:\applics> dirand hit Enter (see Fig. 5.7).42 Chapter Five Figure 5.6 Typical copyright notice and notice provided by the PicBasic compiler when it is run successfully. Figure 5.7 Executing DOS dir (directory) command to see the two additional files (.asm and .hex) created by the PicBasic compiler.Using DOS to Code, Compile, and Program 43 The dir command displays all the files and subdirectories within the subdirectory where it is issued. In Fig. 5.7 we can see the two additional files that the compiler created. One file is the wink.asm file and is the assembler source code file that automatically initiated the macroassembler to compile the assembly code to machine language hexadecimal code. The hex code file is the second file created, called wink.hex. Programming the PIC Chip To program the PIC c hip, we must connect the EPIC programming carrier board (see F ig. 5.8), to the computer. The EPIC board connects to the printer port. The printer port is also called the parallel port.Either name may be used; they are both correct. A computer ma y contain up to four parallel (printer) ports. Each port is assigned a number , from 1 through 4. The computer lists these ports as LPT1 to LPT4. If your computer has only one printer port, disconnect the printer, if one is connected, and attach the EPIC programming board using a 6ft DB25 cable . In some cases it ma y be necessary to temporarily remove the printer driver . Figure 5.9 shows a typical window to disable an HP printer . When you are connecting the programming board to the computer , make sure there are no PIC microcontrollers installed onto the board. If you have an ac adapter for the EPIC Programmer board, plug it into the board. If not, attach two fresh 9V batteries. Connect the “battery on” jumper to apply powFigure 5.8 EPIC programming board and software.44 Chapter Five Figure 5.9 Message box window used to temporarily close printer driver to provide better access to printer port for EPIC programmer. er. The programming board must be connected to the printer port with power applied to the programming board before the softw are is started. If not, the software will not see the programming board connected to the printer port and will give the error message “EPIC Programmer Not Found.” When power is applied and it is connected to the printer port, the LED on the EPIC Programmer board may be on or off at this point.Do not insert a PIC microcontroller into the programming board soc ket until the EPIC program ming software is running. The EPIC Programming Board Software There are two versions of the EPIC softw are: epic.exe for DOS and epicwin.exe for Windows. The Windows softw are is 32bit. It ma y be used with Windows 95, 98, NT, and XP, but not 3.X. It has been my experience that Windows 95 and 98 printer drivers many times like to retain control of the printer (LPT1) port. If this is the case with your computer , the Windows EPIC program may not function properly, and you may be forced to use the DOSlevel program. If you receive an error message “EPIC Programmer Not Found” when you start the EPIC Windows program, youUsing DOS to Code, Compile, and Program 45 have the option of either troubleshooting the problem or using the EPIC DOS program. Using EPIC DOS version If using Windows 95 or higher, you could either open a MSDOS Prompt window or restart the computer in the DOS mode . Windows 3.XX users end the Windows session. Continuing with the wink.bas program Assume we are still in the same DOS session and have just run the PBC compiler on the wink.bas program. We are still in the Applics directory. At the DOS prompt, type EPIC and hit Enter to run the DOS version of the EPIC software (see Fig. 5.10). If you are operating out of a DOS window, you may get a device conflict message box, shown in Fig. 5.11.We want MSDOS to control the LPT port so the EPIC programming softw are will work. Select the MSDOS Prompt and hit the OK button. EPIC’s opening screen is shown in F ig. 5.12. Use the mouse to c lick on the Open button, or press ALTO on your keyboard. Select the wink.hex file (see Fig. 5.13). When the .hex file loads, you will see a list of numbers in the winFigure 5.10 Entering DOS EPIC command to start program.Figure 5.11 Possible “Device Conflict” error message when DOS and Windows both try to use printer port. Select DOS and hit OK button. Figure 5.12 Opening screen of the EPIC programming software. Hit Open File button. 46Using DOS to Code, Compile, and Program 47 Figure 5.13 Select wink.hex in Open File message box and hit the OK button. dow on the left (see F ig. 5.14). This is the mac hine code of your program. On the righthand side of the screen are configuration switc hes that we need to check before we program the PIC c hip. Let’s go through the configuration switc hes once more. Device: Sets the device type. Set it for 8X. Memory Size (K): Sets memory size. Choose 1. OSC: Sets oscillator type. Choose XT for crystal. Watchdog Timer: Choose On. Code Protect: Choose Off. Power Up Timer Enable: Choose On. After the configuration switc hes are set, insert the PIC 16F84 microcontroller into the socket. Click on Program or press ALTP on the keyboard to begin programming. The EPIC program first looks at the microcontroller chip to see if it is blank. If the chip is blank, the EPIC program installs your program into the microcontroller. If the microcontroller is not blank, you are given the options to cancel the operation or overwrite the existing program with the new program. If there is an existing program in the PIC c hip’s memory, write over it.48 Chapter Five Figure 5.14 Hexadecimal numbers showing in EPIC window are the machine language version of the wink.bas program that is uploaded (programmed) into the 16F84 microcontroller. I have noticed that when I place a brand new PICmicro 16F84 c hip into the EPIC compiler to program, EPIC always reports existing code on the c hip. I don’t know if Microchip Technology Inc. loads numbers into the chip’s memory for testing purposes. Don’t let it throw you—the PICmicro c hip is new. The mac hine language code lines are highlighted as the EPIC softw are uploads the program into the PICmicro c hip. When it is finished, the microcontroller is programmed and ready to run. You can verify the program if you like by hitting (or highlighting) the Verify button. This initiates a comparison of the program held in memory to the program stored in the PIC microcontroller.Chapter 6 Testing the PIC Microcontroller The PIC Microcontroller This is where we will build the testing circuit for the PICmicro c hip we programmed. The components needed for the circuit were listed in Chap. 1; if you purchased the components, you can quic kly set up the test circuit. If not, the components are listed again at the end of this c hapter; you will need the components to build the circuit. The solderless breadboard For those of us who ha ve not dabbled in electronics very muc h, I w ant to describe the solderless breadboard (see Fig. 6.1) in detail. As the name implies, you can breadboard (assemble and connect) electronic components onto it without solder. The breadboard is reusable; you can change, modify, or remove circuitry components from the breadboard at any time . This makes it easy to correct any wiring errors . The solderless breadboard is an important item for constructing and testing circuits outlined in this book. The style of breadboard on the left is a vailable from any number of sources including RadioShack. The breadboard on the right is similar but provides a larger prototyping area. If you wish to make any circuit permanent,you can transfer the components onto a standard printedcircuit board and solder it together with the foreknowledge that the circuit functions properly. A partial cutaway of the top surface shows some of the internal structure of a board (F ig. 6.2). The holes on the board are plugs . When a wire or pin is inserted into the hole, it makes intimate contact with the metal connector strip inside. The holes are properly distanced so that integrated circuits and many other components can be plugged in. You connect components on the board by Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 4950 Chapter Six Figure 6.1 Top view of solderless breadboards. Figure 6.2 Top view of solderless breadboards with a partial cutaway showing underneath conductive contact strips. using 22gauge (solid or stranded) wire . I prefer to use stranded wire because it has greater flexibility; other people prefer solid wire because it’ s stiffer and easier to push into the breadboard hole . The complete internal wiring structure of the solderless boards is shown in Fig. 6.3. The solderless breadboard on the left shows the X and Y rows that are typically used to supply power (Vcc) and ground connections to the circuit.The columns below the X row and above the Y row are used for mounting components. The solderless breadboard on the right has double rows located at the top and bottom. These are used to supply both Vcc and ground on each side of the breadboard. Three schematics, one circuit Figures 6.4, 6.5, and 6.6 are identical schematics of our test circuit. The 16F84 PIC microcontroller in the sc hematic is the microcontroller you programmed in either Chap . 4 or 5. I drew three sc hematics to help orient experimenters who may not be familiar with standard electrical dra wings. Figure 6.4 showsTesting the PIC Microcontroller 51 Figure 6.3 Top view of solderless breadboards detailing conductive strips . + Electrical Symbol Component Appearance 4.0 MHz 470 470 4.7kW +5V 123456789 18 17 16 15 14 13 12 11 10 RA2 RA1 RA0 OSC1/CLKIN OSC2/CLKOUT Vdd RB7 RB6 RB5 RB4 RA3 RA4IT0CKI MCLR RB0/INT RB1 RB2 RB3 Vss PIC16F84 22pF 22pF 4MHz .1uF Xtal LED Resistor Capacitor All resistors 1/4 watt + + Figure 6.4 Isometric schematic of test circuit for wink.bas program. how the PIC 16F84 microcontroller and components appear. There is a legend at the bottom that shows the electrical symbol and the typical appearance of the component. Figure 6.5 is a line dra wing showing how the components appear mounted on one of the solderless breadboards . The writing on Fig. 6.5 points out each electrical component.52 Chapter Six ABCDE FGHIJ YX 11 55 10 10 15 15 20 20 PIC16F84 7805 Volt Reg. Xtal 22 pF LEDs Red LED Side View + ++ 470W Ground 4.7kW Ground +Vcc Gnd .1uF Figure 6.5 Isometric drawing showing test circuit constructed on solderless breadboard. Figure 6.6 Schematic of test circuit for wink.bas program. If you examine the placement of the components mounted on the solderless breadboard with its internal electrical wiring (F igs. 6.2 and 6.3), you can see how the components connect to one another and produce a circuit. Figure 6.6 is the same sc hematic drawn as a standard electrical dra wing with the pin numbers grouped and oriented to function. For the remainder of the book, standard electrical drawings will be used. The schematic shows how minimal are the components needed to get your microcontroller up and running . Primarily you need a pullup resistor on pinTesting the PIC Microcontroller 53 Figure 6.7 Photograph of wink.bas circuit constructed on solderless breadboard. 4 (MCLR), a 4MHz crystal with two (22pF) capacitors and a 5V power supply. Note: The 4MHz crystal and two (22pF) capacitors make up an oscillator that is required by the microcontroller . These three parts ma y be substituted with a 4MHz ceramic resonator. The two LEDs and the two resistors connected in series with eac h LED are the output. It allows us to see that the microcontroller and program are functioning properly. Assemble the components as shown in the sc hematic (Fig. 6.5) onto the solderless breadboard. When you ha ve finished, your work should appear as in Fig. 6.7. Although the specifications sheet on the 16F84 states the microcontroller will operate on voltages from 2 to 6 V, I provided a regulated 5V power supply for the circuit. The regulated power supply consists of a 7805 voltage regulator and two filter capacitors. Wink Apply power to the circuit. The LEDs connected to the c hip will alternately turn on and off.Wink, …, wink. Now you know how easy it is to program these microcontrollers and get them up and running .54 Chapter Six Troubleshooting the circuit There is not too muc h that can go wrong here . If the LEDs do not light, the first thing to check is the orientation of the LEDs. If they are put in backward, they will not light. Next check your ground wires. See the jumper wires on the righthand side of the solderless breadboard.They bring the ground up to the two 22pF capacitors. Check all your connections . Look back at F igs. 6.2 and 6.3 to see how the underlying conductive strips relate to the push in terminals on top of the board. PIC Experimenter’s Board and LCD Display There are two optional tools you ma y want if you plan on experimenting with the PIC16F84 and microcontrollers in general. They are the PIC Experimenter’s Board and LCD displa y. We will look at the LCD displa y first because a similar LCD displa y is incorporated into the PIC Experimenter’ s Board and what we say about the standalone LCD display is also true for the PIC Experimenter’s Board LCD display. One thing PIC microcontrollers lac k is some type of displa y.With a display, the chip could show us how a program is running or what it is detecting . In addition a displa y would allow the microcontroller to output textual and numeric messages to the user. To this end there are serial LCD displays on the market that only require a single microcontroller’ s I/O lines (pin) and a circuit ground. The particular LCD display we are using receives standard serial data (RS232) at 300, 1200, 2400, and 9600 baud (Bd) (inverted or true).The LCD module is a twoline, 16character visible display. The full display is actually two lines by 40 characters, but the additional 24 c haracters per line are off screen. We can use the PicBasic and PicBasic Pro serout command to communicate and output messages to the LCD display. The PicBasic and PicBasic Pro compilers can send and receive serial information at 300, 1200, 2400, and 9600 Bd. Data are sent as 8 bits , no parity, and 1 stop bit. The serial mode ma y be set to be true or inverted. These data match the serial communication protocols required of the LCD display. The LCD module has three wires: 15 V (red), GND (black or brown), and a serial in line (white). The baud rate may be set to 300, 1200, 2400, or 9600 by using a set of jumpers (J1, J2, and J3) on the back of the LCD display. This first program prints the message “Hello World.” The cursor (printing position) automatically moves from left to right.The schematic is shown in Fig. 6.8, and the LCD display is shown in Fig. 6.9. ‘PicBasic program‘LCD testpause 1000 ‘Wait 1 second for LCD to initializeTesting the PIC Microcontroller 55 PIC Microcontroller Ground Serial Line +5V Figure 6.8 Schematic of LCD display test circuit. Figure 6.9 Photograph of LCD display “Hello World.” start: serout 1, t1200, (254,1) ‘Clear screen pause 40 serout 1, t1200, (“Hello World”) ‘Print message pause 400 goto start end I kept this program small to show how easy it is to print a message on the LCD display. Here is the same program written for the PicBasic Pro compiler . ‘PicBasic Pro program‘LCD testpause 1000 ‘Wait 1 second for LCD to initializestart:serout portb.1, 1, [254,1] ‘Clear screenpause 40serout portb.1, 1, [“Hello World”] ‘Print messagepause 400goto startend56 Chapter SixNotice that, in line 5 of the program(s),serout 1, t1200, (254,1)is a command. The LCD module has eight common commands . All commands are prefixed with the decimal number 254. The LCD module will interpret any number following a 254 prefix as an instruction. Instead of decimal numbers , you may also use hexadecimal numbers , if you wish. So in hexadecimal the command becomes serout 1, t1200, ($fe, $01). The following is a list of a few common commands. Remember all commands are prefixed with a 254 ($fe). Code Instruction 1 Clear screen. 2 Home position (move cursor top left of displa y). 16 Move cursor one character position left. 20 Move cursor one character position right. 24 Scroll display one character position left. 28 Scroll display one character position right. 192 Move cursor to first position on second line . PIC Experimenter’s Board The PIC Experimenter’s Board is a prefabricated developing board for prototyping circuits (see Fig. 6.10). The board allows easy access to all the I/O pins, port A (RA0–RA4), and port B (RB0–RB7) of the 16F84.The board may also be used with the 16F8X, 16C55X, 16C62X, 16C7X, and 16C8X family of 18pin PIC microcontrollers. Its 168point solderless connection area allows for quic k and easy access to all port A (RA0–RA4) and port B (RB0–RB7) I/O lines. There is an open 18pin socket for inserting the microcontroller you are developing. The board includes an integrated 16 3 2 serial LCD display (optional backlight), which can be easily connected with one wire to any I/O line (or external source). Use The board can be powered by either an onboard 9V battery or an ac/dc trans former. The power switch in the upper right turns power to the board on and off. The board inc ludes a reset button, for resetting the microcontroller . The LCD has its own power switc h, located directly above the LCD . If your LCD has a backlight, the backlight switch is located above the LCD power switc h. I will describe the prototyping section on the PIC Experimenter’ s Board, as I did with the solderless breadboards , and finish up the description by wiring a simple microcontroller LED project on the Experimenter’s Board. The prototyping is located at the lower left corner of the PIC Experimenter’s Board (see Fig. 6.11). There is an open 18pin soc ket to hold the microcontroller being developed.Testing the PIC Microcontroller 57 Figure 6.10 Photograph of PIC Experimenter’s prototype developing board. The prototyping area is similar in design and function to solderless bread boards; see Fig. 6.12. You can breadboard (assemble and connect) electronic components and electronic circuits into the prototyping area without solder ing. The prototyping area is reusable; you can c hange, modify, or remove circuit components at any time. This makes it easy to correct any wiring errors . A cutaway of the prototyping area is shown in F ig. 6.13. The square holes shown in the area are soc kets. When a wire or pin is inserted into a hole , it makes electrical contact with the underlying metal strip . The holes are spaced so that integrated circuits and many other components can be plugged right in. The internal electrical connection structure of the prototyping area is shown in Fig. 6.14. Looking at F ig. 6.15, at the top of the prototyping area we see that the columns of bank 1 are labeled with the pin assignments from the 16F84.These columns are directly connected to those microcontroller pins . Connecting a wire or device to any of the three sockets in a particular column is electrically connecting that wire or device to that I/O pin of the 16F84. Bank 2 provides 14 individual foursocket columns. The four sockets aligned in each individual column are electrically connected. The individual columns are separate electrically from one another. Bank 3 is the same as bank 2.58 Chapter SixFigure 6.11 Photograph of PIC Experimenter’s Board with breadboarding area and 18pin socket highlighted. Figure 6.12 Diagram of the breadboard area. The last row, labeled GND (ground), is electrically connected across the entire row. There are an additional three ground soc kets at the top of bank 1. A 15V power is available from a foursocket column adjacent to bank 1. Simple experiment We shall wire a simple experiment to illustrate the use of the experimenter’ s prototyping area: blinking an LED . Yes, this is very similar to the wink proTesting the PIC Microcontroller 59 Figure 6.13 Cutaway view of breadboard area. Figure 6.14 Underlying electrical connections of breadboard area. gram, with the exception that we are only using one LED this time . The following are a small PicBasic program and PicBasic Pro program to blink an LED on pin RB1. PicBasic program PicBasic Pro program start: high 1 start: high portb.1pause 250 pause 25060 Chapter SixFigure 6.15 Diagram of breadboard area with banks, ground, and15V power supply highlighted.low 1 low portb.1pause 250 pause 250goto start goto startThe complete schematic for this experiment is shown in Fig. 6.16. Aside from a programmed 16F84, we only need two other components: a 470V, 1/4W resistor and a subminiature LED . All the other components needed to make the 16F84 work are already hardwired on the PIC Experimenter’ s Board. The LED has two terminals, one longer than the other. The longer terminal on the LED is positive, shown in the legend of Fig. 6.17. On the schematic the LED appears as a diode. To wire this circuit, connect one lead of the 1/4W resistor into one of the RB1 sockets. Connect the other lead of the 1/4W resistor into a socket in bank 2. Take the positive lead of the LED and plug it into a soc ket in the same column as the one containing the resistor lead. Connect the opposite lead of the LED, and plug it into one of the ground soc kets at the bottom. Plug the programmed 16F84 microcontroller into the 18pin soc ket on the PIC Experimenter’ s board, and turn on the power . The LED should begin blinking on for 1/4 s, then off for a 1/4 s. This on/off cycle (blinking) continually repeats. Builtin LCD The onboard LCD display combines a serial interface and a 2line by 16c haracter display (see Fig. 6.18). The LCD display can be set to receive serial data at 300, 1200, or 2400 Bd (true or inverted, switchselectable).Testing the PIC Microcontroller 61 Figure 6.16 Schematic of blink circuit. Figure 6.17 Diagram of blink circuit assembled in breadboard area of PIC Experimenter’s Board. To use the LCD, connect a jumper from the desired output pin on the microcontroller to the serial input. It is not necessary to connect a secondary ground line to the serial input ground unless the serial data are coming from a source off the Experimenter’s Board. Like the LCD module, the onboard LCD display has two operational modes: text and instruction. The default is text mode; data received via the serial input line appears on the screen. Send the string “Images” and “Images” will appear on the LCD . To input instructions to the LCD displa y, such as c lear62 Chapter SixFigure 6.18 Photograph of PIC Experimenter’s Board with LCD controls highlighted. screen, go to line 2, etc., you must prefix the instruction with ASCII 254 (0xFE). The byte following the prefix is seen and treated as an instruction code. After the instruction code , the unit automatically returns to text mode . Every instruction code must be sent with its own 254 prefix. If your LCD is backlit, you may adjust the backlight contrast to the optimal setting via the LCD contrast control. The contrast control is set fully clockwise (highest contrast) at the factory, but you can adjust the control by hand. To set the baud rate , there are three sets of jumpers: J1 to J3. Set the jumpers in accordance with the silkscreen diagram on the Experimenter’ s PC Board. At all baud rates , serial data are received at 8 data bits , 1 stop bit, no parity. Note that the baud rate setting is only read once at startup , so changing the jumpers while the module is active will not have any effect on the baud rate until the Experimenter’s Board is reset. Once the LCD module is connected and configured to matc h the baud rate of the computer/microcontroller, it will receive those transmitted data and display the information on the LCD display. For example, if you send “Hello,” then “Hello” appears on the displa y. The cursor (printing position) automatically moves from left to right. The onboard LCD display will accept the standard LCD instructions. A particular byte is identified as an instruction when it is preceded by an instruction prefix c haracter, ASCII 254 (0xFE hex, 11111110 binary). The onboard LCD treats the byte immediately after the prefix as an instruction, then automatically returns to data mode . For example, the c learscreen instruction is ASCII 1. To clear the screen, send ,254.,1. (where the , . symbols mean single bytes set to these values , not text as typed from the keyboard). Notice this instruction code matc hes the instruction code for the serial LCD displa y module. Instruction Code (decimal) Clear screen 1Testing the PIC Microcontroller 63 Figure 6.19 Photograph of LCD display in selftest mode. Home position (move cursor top left of displa y) 2 Move cursor one character position left 16 Move cursor one character position right 20 Scroll display one character position left 24 Scroll display one character position right 28 Set cursor position (DDRAM address) 128 1 addr Set point in character generator (CG) RAM 64 1 addr The onboard LCD also has a selftesting mode that will print the current baud rate as determined by the jumper settings and mode (true/inverted); see Fig. 6.19. To enter selftest mode, connect the serial in line to ground (for true) or 15 V (for inverted) upon LCD module startup . Note: If the serial input line is improperly connected for selftest mode , for instance connected to 15 V when jumpers are set for true mode , the LCD display will remain blank. The module stays in selftest mode as long as the serial input line is held either high (inverted mode) or low (true mode). LCD module may be exited from selftest mode on the fly by simply connecting the serial input line to a serial source . When you print past the end of a line , the next 24 characters do not show up on the LCD screen. They are not lost; they are in an offscreen memory area. All alphanumeric LCD modules ha ve 80 bytes of memory , arranged appropriately for a 2 3 40 screen. On LCDs with smaller screens (suc h as this 2 3 16), text printed past the end of a visible line goes into memory, but can’t be seen on the screen. Use cursorpositioning instructions to print to a particular location on the displa y. Or deliberately print in offscreen memory to temporarily hide text, then send scrollleft instructions to reveal it. Using the LCD: PicBasic and PicBasic Pro examples Connect the serial input of the LCD to portb .0 of a PIC microcontroller. The following PicBasic program demonstrates sending data and commands to the LCD. main: pause 1000 ‘Wait for the LCD to start upserout 0, t2400, ($fe,$01) ‘Clear the screen64 Chapter Six pause 40 serout 0, t2400, (“Wherever you go”) serout 0, t2400, ($fe,$c0) ‘Move to line 2 serout 0, t2400, (“there you are”) pause 1000 ‘Wait 1 second goto main ‘Do it again The program clears the LCD and sends the message “Wherever you go there you are” at 2400 Bd (true mode), waits for 2 s, and then loops indefinitely (see Fig. 6.20). Note that in the above example , the control codes were written in hexadecimal (base16). Hexadecimal is specified by a dollar sign ($) prefix. If you wish, the lines serout 0, t2400,($fe,$01) ‘Clear the screenserout 0, t2400,($fe,$c0) ‘Move to line 2could also be written with decimal (base10) notation as serout 0, t2400,(254,1) ‘Clear the screenserout 0, t2400,(254,192) ‘Move to line 2Which notation you choose to use is a matter of preference. An equivalent program can be written with PicBasic Pro as follows: Figure 6.20 Photograph of message from pin RB0 of 16F84 microcontroller sent to onboard LCD display.Testing the PIC Microcontroller 65 main: pause 1000 ‘Wait for the LCD to start up serout portb.0,0,[$fe,$01] ‘Clear the screen pause 40 serout portb.0,0,[“Wherever you go”] serout portb.0,0,[$fe,$c0] ‘Move to line 2 serout portb.0,0,[“there you are”] pause 1000 ‘Pause for a second goto main ‘Loop Introduction to Binary and the PIC Microcontroller The term binary means “based on 2,” as in two numbers 0 and 1. It’s also like an electric switch that has two values: on (1) and off (0). The term bit is an acronym that stands for the term binary digit. A bit or binary digit can ha ve two values , either 0 or 1. A byte is a digital expression (number) containing 8 bits. Binary is important to computers and microcontrollers . The bit values of 0 and 1 are the only things a computer can read.Actually the computer or microcontroller can’t really read, but it can sense voltage values . So a bit that is on 1 is represented by a positive voltage . Consequentially a bit 0 is off 0 and is represented as no voltage. A single bit by itself is of little value , but start putting them together to make bytes (8 bits), words (16 bits, 32 bits, 64 bits, 128 bits), and so on, and we can make the computers perform mathematics , create word processors and spreadsheets, create a cyberspace (Internet), etc. All these applications are based on a bit. To read or write to a port register requires understanding a little binary . When we read and write to any port, we use standard decimal numbers . However it’ s the binary equivalent of those decimal numbers that the PIC microcontroller sees and uses. The 16F84 uses 8bit port registers so we only need to concern ourselves with small 8bit numbers and their decimal equivalents. Remember that an 8bit number is called a byte . An 8bit number can represent any decimal value between 0 and 255. When we write a decimal number into a register , the PIC microcontroller can only see the binary equivalent of that decimal number (byte) we wrote to the register . For us to understand what’ s happening inside the register, we need to be able to look at the binary equivalents of the decimal (byte) number also. Once we can do this, our ability to effectively and elegantly program the PIC microcontrollers is greatly enhanced. Examine the binary number table at the top of the next page. It shows all the decimal and binary number equivalents for numbers 0 through 32. By using this information, the binary numbers from 32 to 255 can be extrapolated. Each decimal number on the left side of the equals sign has its binary equivalent on the right side. So where we see a decimal number, the microcontroller will see the same number as a series of 8 bits (there are 8 bits to a byte).66 Chapter SixBinary Number Table 0 5 00000000 16 5 00010000 32 5 00100000 1 5 00000001 17 5 00010001 . 2 5 00000010 18 5 00010010 . 3 5 00000011 19 5 00010011 . 4 5 00000100 20 5 00010100 64 5 01000000 5 5 00000101 21 5 00010101 . 6 5 00000110 22 5 00010110 . 7 5 00000111 23 5 00010111 . 8 5 00001000 24 5 00011000 128 5 10000000 9 5 00001001 25 5 00011001 . 10 5 00001010 26 5 00011010 . 11 5 00001011 27 5 00011011 . 12 5 00001100 28 5 00011100 255 5 11111111 13 5 00001101 29 5 00011101 14 5 00001110 30 5 00011110 15 5 00001111 31 5 00011111 Figure 6.21 shows the relationship between a binary number and the two PIC microcontroller registers that control port b. Notice each register has eight open positions. This register can hold an 8bit (1byte) number. Let’s look at the second binary number table below. Notice for each progression of the binary 1 to the left, the exponential power of 2 is increased by 1. Second Binary Number Table Bit no. Decimal Binary Bit no. Decimal Binary Bit 0 1 5 00000001 Bit 4 16 5 00010000 Bit 1 2 5 00000010 Bit 5 32 5 00100000 Bit 2 4 5 00000100 Bit 6 64 5 01000000 Bit 3 8 5 00001000 Bit 7 128 5 10000000 These are relevant numbers , because each progression to the left identifies another bit location and bit weight within the 8bit byte . For instance , suppose we w anted to write binary 1s at the RB6 and RB2 locations. To do so, we add their bit weights, in this case 64 (RB6) plus 4 (RB2), which equals 68. The binary equivalent of decimal number 68 is 01000100. If you push that number into the port B register, you will see that the binary 1s are in the RB6 and RB2 positions . Remember this—it is important. The open TRISB register shown in F ig. 6.21 may be used to examine numbers placed in the TRISB. The port B register ma y be used to examine num bers placed at the port B register. Notice the correlation between the register bit locations , bit weights, andTRISB Decimal 134 86 Hex Port B Port B Decimal 6 06 Hex Binary 00000001 00000010 00000100 00001000 00010000 00100000 01000000 10000000 Power of Two 20 = 1 21 = 2 22 = 4 23 = 8 24 = 16 25 = 32 26 = 64 27 = 128 Bit Weight/Values Register Location 128 64 32 16 8 4 2 1 RB0 RB1 RB2 RB3 RB4 RB5 RB6 RB7 Binary 00000001 00000010 00000100 00001000 00010000 00100000 01000000 10000000 Power of Two 20 = 1 21 = 2 22 = 4 23 = 8 24 = 16 25 = 32 26 = 64 27 = 128 Bit Weight/Values Register Location 128 64 32 16 8 4 2 1 RB0 RB1 RB2 RB3 RB4 RB5 RB6 RB7 Figure 6.21 Diagram of port B registers. 6768 Chapter Sixport B I/O pins. This correspondence between the bit number , bit weight, and the I/O line is used to program and control the port. Using the TRIS and port registers The TRIS (tristate enable) register is a 1byte (8bit) programmable register on the PIC 16F84 that controls whether a particular I/O pin is configured as an input or output pin. There is a TRIS register for each port. TRISA controls the I/O status for the pins on port A, and TRISB controls the I/O status for the pins on port B. If one places a binary 0 at a bit location in TRISB for port B, the corresponding pin location on port B will become an output pin. If one places a binary 1 at a bit location in the TRISB, the corresponding pin on port B becomes an input pin. The TRISB data memory address for port B is 134 (or 86h in hex). After port B has been configured using the TRISB register, the user can read or write to the port, using a port B address (decimal number 6). Here is an example. Suppose we want to make all port B lines output lines . To do so, we need to put a binary 0 in eac h bit position in the TRISB register. So the number we would write into the register is decimal 0. Now all our I/O lines are configured as output lines . If we connect an LED to eac h output line, we can see a visual indication of any number we write to port B . If we want to turn on the LEDs connected to RB2 and RB6, we need to place a binary 1 at eac h bit position on port B register. To accomplish this, we look at the bit weights associated with eac h line. RB2 has a bit weight of 4, and RB6 has a bit weight of 64. We add these numbers (4 1 64 5 68) and write that number into the port B register . When we write the number 68 into the port B register , the LEDs connected to RB2 and RB6 will light. To configure port A, we use the TRISA register, decimal address 133 (see Fig. 6.22). On port A, however, only the first 5 bits of the TRISA and the corresponding I/O lines (RA0–RA4) are a vailable for use. Examine the I/O pinout on the 16F84, and you will find there are only five I/O pins (RA0–RA4) corre sponding to port A. These pins are configured using the TRISA register and used with the port A address. Memory location, Memory location, Register hexadecimal decimal Port A 05h 5 Port B 06h 6 TRISA 85h 133 TRISB 86h 134 On power up and reset, all the I/O pins of port B and port A are initialized (configured) as input pins . We can change this configuration with our program.TRISA Decimal 133 85 Hex Port A Port A Decimal 5 05 Hex Binary 00000001 00000010 00000100 00001000 00010000 Power of Two 20 = 1 21 = 2 22 = 4 23 = 8 24 = 16 Bit Weight/Values Register Location 16 8 4 2 1 RA0 RA1 RA2 RA3 RA4 Binary 00000001 00000010 00000100 00001000 00010000 Power of Two 20 = 1 21 = 2 22 = 4 23 = 8 24 = 16 Bit Weight/Values Register Location 16 8 4 2 1 RA0 RA1 RA2 RA3 RA4 Figure 6.22 Diagram of port A registers. 6970 Chapter Six Here’s another example. Let’s configure port B so that bit 0 (RB0) is an input pin and all other pins are output lines . To place binary 0s and 1 in the proper bit location, we use the bit weights shown in the binary number table . For instance, to turn bit 0 on (1) and all other bits off (0), we would write the decimal number 1 into TRISB for port B. Depending upon which PicBasic compiler is used, the commands are a little different. For the PicBasic compiler, the command to write to a register is the poke command. The program line to write the decimal value 1 into the TRISB register will look like this: poke 134,1The number after the poke command is the memory address that the command will write to, in this case 134. The number 134 is the memory address of the TRISB for port B. The next number, separated by a comma, is the value we want to write in that memory address . In this case it’s the number 1. For the PicBasic Pro compiler , the TRISB and TRISA registers are already predefined. Thus when the compiler sees TRISB, it accesses the proper memory (134) location. So the equivalent command for the PicBasic Pro is TRISB = 1Look at the binary equivalent of the decimal number 1: 0 0 0 0 0 0 0 1 Mentally place eac h 1 and 0 into the TRISB register locations shown in F ig. 6.21. See how the 1 fits into the bit 0 place, making that corresponding line an input line, while all other bit locations have a 0 written in them, making them output lines. So by poking (writing) this location with a decimal number that represents a binary number containing the proper sequence of bits (0s and 1s), we can configure any pin in the port to be either an output or an input in any combi nation we might require . In addition, we can c hange the configuration of the port “on the fly” as the program is running. To summarize, writing a binary 1 into the TRIS register turns that corresponding bit/pin on the port to an input pin. Likewise, poking a binary 0 will turn the bit into an output. Accessing the ports for output Once the port lines have been configured (input or output) using the TRIS register, we can start using it.To output a binary number at the port,simply write the number to the port, using the poke (PicBasic) or trisx.x (PicBasic Pro) command. The binary equivalent of the decimal number will be outputted, as shown in our first example. To output a high signal on RB3 using the PicBasic compiler, we could use this command:Testing the PIC Microcontroller 71 Figure 6.23 Schematic of eight LEDs connected to port B for counting program. poke 6, 8where 6 is the memory address for port B and 8 is the decimal equivalent of the binary number (00001000) we want to output. For the PicBasic Pro compiler, the equivalent command is output portb.3 = 1Counting program To illustrate many of these concepts , I have written a simple basic program. The schematic for the program is shown in F ig. 6.23. It is a binary counting program that will light eight LEDs connected to port B’ s eight output lines. The counting program will light the LEDs in the sequence shown in the binary number table. Each binary 1 in a number the table will be represented with a lit LED. Every 250 milliseconds (ms) (1/4 s), the count increments. After reaching the binary number 255 (the maximum value of a byte), the sequence repeats, starting from zero. Counting in binary by 1 The following program is written for the PicBasic compiler . ‘Program binary counting‘Initialize variables symbol trisb = 134 ‘Assign TRISB of port b to decimal value of 134 symbol portb = 6 ‘Assign port b to decimal value of 6 ‘Initialize port(s) poke trisb,0 ‘Set port b pins to output loop:72 Chapter Six for b0 = 0 to 255 poke portb, b0 ‘Place count at port b to light LEDs pause 250 ‘Pause 1/4 s or it’s too fast to see next b0 ‘Next counter value goto loop ‘Start over again ‘End The following program is written for the PicBasic Pro compiler . ‘Program binary counting‘Initialize variables ct var byte ‘Counting variable ‘Initialize port trisb = 0 ‘Set port b pins to output loop: for ct = 0 to 255 ‘Counter portb = ct ‘Place counter on port b to light LEDs pause 250 ‘Pause 1/4 s next ct ‘Next counter value goto loop ‘Start over again ‘End Input The ability of our microcontroller to read the electrical status of its pin(s) allows the microcontroller to see the outside world. The line (pin) status ma y represent a switc h, sensor, or electrical information from another circuit or computer. The button command The PicBasic compiler comes equipped with a simple command to read the electrical status of a pin, called the button command. The button command, while useful, has a few limitations . One limitation of this command is that it may only be used with the eight pins that make up port B. The I/O pins available on port A cannot be read with the button command. Another limitation is that you cannot read multiple port pin inputs at once, only one pin at a time. We will overcome these button command limitations using the peek command. But for the time being, let’s use and understand the button command. As the name implies , the button command is made to read the status of an electrical button switch connected to a port B pin. Figure 6.24 shows two basic switch schematics, labeled A and B, of a simple switch connected to an I/O pin. The button command structure is as follows: button pin, down, delay, rate, var, action, labelTesting the PIC Microcontroller 73 Figure 6.24 Schematic of electric switches suitable for use with PIC microcontrollers. Pin Pin number (0–7), port B. Down State of pin when button is pressed (0 or 1). Delay Cycle count before auto repeat starts (0–255). If 0, no debounce or autorepeat is performed. If 255, debounce, but no autorepeat is performed. Rate Autorepeat rate (0–255). Var Byte variable used for delay/repeat countdown. Should be initialized to 0 prior to use. Action State of button to perform goto (0 if not pressed, 1 if pressed). Label Execution resumes at this label if Action is true. Let’s take another look at the switc h schematic in Fig. 6.24 before we start using the button switc h. Let’s visualize how the switc hes affect the I/O pin electrically. The switch labeled A in Fig. 6.24 connects the I/O pin to a 15V power supply through a 10,000 V resistor. With the switch open, the electrical status of the I/O pin is kept high (binary 1). When the switch is closed, the I/O pin connects to ground, and the status of the I/O pin is brought low (binary 0). The switc h labeled B in F ig. 6.24 has an electrical function opposite the switch labeled A. In this case, when the switch is open, the I/O pin is connected to ground, keeping the I/O pin low (binary 0).When the switch is closed, the I/O pin is brought high (binary 1). In place of a switc h, we can substitute an electric signal, high or low, that can also be read using the button command. Typically the button command is used inside a program loop , where the program is looking for a change of state (switch closure). When the state of the I/O pin (line) matc hes the state defined in the Down parameter , the program execution jumps out of the loop to the label portion of the program.74 Chapter Six A button example If we want to read the status of a switc h of I/O pin 7, here is a command we will use in the next program. button 7, 0,254,0,b1,1,loopThe next program is similar to the previous program 3, inasmuch as it performs a binary counting. However, since we are using PB7 (pin 7) as an input, and not an output, we lose its bit weight in the number we can output to port B. The bit weight for pin 7 is 128. So without pin 7 we can only displa y numbers up to decimal number 127 (255 2 128 5 127). This is reflected in the first loop (pin7/bit 7 5 128). The program contains two loops . The first loop counts to 127, and the current number’s binary equivalent is reflected by the Lite LEDs connected to port B. The loop continues to count as long as the switc h SW1 remains open. When SW1 is closed, the button command jumps out of loop 1 into loop 2. Loop 2 is a noncounting loop where the program remains until switc h SW1 is reopened. You can switc h back and forth between counting and noncounting states. Figure 6.25 is a schematic of our button test circuit. The following program is written for the PicBasic compiler . ‘Program for PicBasic compilersymbol trisb = 134 ‘Set TRISB to 134 symbol portb = 6 ‘Set port b to 6 ‘Initialize Port(s) poke trisb,128 ‘Set port b pins (16 output), pin 7 input Figure 6.25 Schematic of seven LEDs and one switch connected to port B for the switch detection and counting program.Testing the PIC Microcontroller 75 label 1: b1s = 0 ‘Set button variable to 0 loop1: ‘Counting loop for b0 = 0 to 127 poke portb, b0 ‘Place b0 value at port to light LEDs pause 250 ‘Pause counting or it’s too fast to see button 7,0,254,0,b1,1,label2 ‘Check button status; if closed, jump next b0 ‘Next b0 value goto loop1 label12: b1=0 ‘Set button variable to 0 loop2: ‘Second loop not counting poke portb,0 ‘Turn off all LEDs button 7,1,254,0,b1,1,label1 ‘Check button status; if open, jump back goto loop2 When the program is run, it begins counting. When the switch is closed, all the LEDs will turn off , and it stops counting . Open the switc h, and the counting resumes, starting from 0. ‘Program for PicBasic Pro compilerct var bytec1 var byte‘Initialize port(s)trisb = 128label1: c1=0 loop1: For ct = 0 to 127portb = ct pause 250 button 7,0,254,0,c1,1,label2 next ct goto loop1label2: c1=0 loop2: portb =0 button 7,1,254,0,c1,1,label1goto loop2peek ‘Set port b pins (16 output), pin 7 input‘Set button variable to 0‘Counting loop‘Place ct value at port to light LEDs‘Pause counting or it’s too fast to see‘Check button status; if closed, jump‘Next counting value‘Set button variable to 0‘Second loop not counting‘Turn off all LEDs‘Check button status; if open, jump backThe peek command can only be used with the PicBasic compiler .We can also use the peek command to check the status of any input line . The advantages of the peek command are as follows . Using peek, we can read the five I/O lines of port A (or the eight I/O lines of port B) at once. This increases the versatility of the PIC c hip and allows our program to be more concise (less convoluted), shorter, and easier to read. To emphasize these points , let’s rewrite our last programs , using the peek command. This program uses the same sc hematic.76 Chapter Six ‘PicBasic program that uses the peek command‘Initialize port(s) symbol trisb = 134 ‘Set TRISB to 134 symbol portb = 6 ‘Set port b to 6 poke trisb,128 ‘Set port b pins (1..6) output, pin 7 input loop1: ‘Counting loop for b0 = 0 to 127 poke portb, b0 ‘Place b0 value at port to light LEDs pause 250 ‘Pause counting or it’s too fast to see peek portb,b0 ‘Check button status if bit7 = 0 then loop2 ‘If sw1 is closed, jump to loop2 next b0 ‘Next b0 value goto loop1 loop2: ‘Second loop not counting pke portb,0 ‘Turn off all LEDs peek portb,b0 ‘Check button status; if open, jump back if bit7 = 1 then loop1 ‘If sw1 is open, jump to loop 1 goto loop2 The variable b0 is performing double duty . First it is holding our current counting numbers 0 through 127. The numbers 0 to 127 require 7 bits of the variable b0 (bit 0 through bit 6). This leaves the eighth bit (bit 7) available for use.We use bit 7 to check the status of the switch. If it’s open, its value will be a binary 1; if it’s closed, it’s equal to binary 0. The command peek is followed by a memory address, then a comma, then a storage variable. peek address, varAs its name implies , the peek command allows one to view (or peek at) the contents of a specified memory address. Typically the memory address “peeked at” is one of the PIC microcontroller’s registers. The “peeked” value is stored in a variable var defined in the command. In this program we peeked at the one input line on port B: peek portb,b0The peek command can read an entire byte (8 bits) at once . Or as in this case, only the upper bit (bit 7) of the peeked value is relevant. (The rest of the bits are holding our counting number that’ s outputted to the LEDs). peek and PicBasic Pro When you are using the PicBasic Pro compiler , it is recommended not to use the peek command. Fortunately there is an easy workaround to the peek command.We simply assign a variable to the port we wish to peek at. var = portbTesting the PIC Microcontroller 77 The value placed in the variable var is our peek value. ‘PicBasic Pro program that uses a peek equivalent command ct var byte c1 var byte ‘Initialize port(s) trisb = 128 ‘Set port b pins (1..6) output, pin 7 input loop1: ‘Counting loop for ct = 0 to 127 portb = ct ‘Place ct value at port to light LEDs pause 250 ‘Pause counting or it’s too fast to see c1 = portb ‘Check button status if c1.7 = 0 then loop2 ‘If sw1 is closed, jump to loop2 next ct ‘Next ct value goto loop1 loop2: ‘Second loop not counting portb = 0 ‘Turn off all LEDs c1 = portb ‘Check button status; if open, jump back if c1.7 = 1 then loop1 ‘If sw1 is open, jump to loop 1 goto loop2 Basic input and output commands In our programs we directly wrote to the PIC microcontroller TRIS registers (A or B) and port registers . By doing so we were able to create input and out put pins and then access them in our programs . There are other commands you can use to accomplish the same thing . The PicBasic and PicBasic Pro compilers have two basic commands for making individual pins either input or output lines. The commands are input and output. Unfortunately these two basic commands only work on port B pins (0 to 7) for PicBasic. For PicBasic Pro, any port may be used. The command input pinmakes the specified pin an input line . Only the pin number itself, that is, 0 to 7, is specified (i.e., not pin 0), for example, input 2 ‘Makes pin2 (rb2) an input line.The opposite of the input command is the output command. The command output pinmakes the specified pin an output line. Only the pin number itself, that is, 0 to 7, is specified (i.e., not pin 0), for example, output 0 ‘Makes port b, pin 0 (rb0) an outputThe above examples are intended for use with either PicBasic or PicBasic Pro.78 Chapter Six The PicBasic Pro has an additional command structure that can be used with both the input and output commands. This allows one to make input and output pins on other ports besides port B . This is accomplished by specifying the port and the pin. For instance, to access port A, pin 2, you use the following format: porta.2To use this in a command: input porta.2 ‘Make port a, pin 2 an inputoutput porta.3 ‘Make port a, pin 3 an outputServomotorsServomotors (see Fig. 6.26) are used in many radiocontrolled model airplanes, cars, boats, and helicopters. Because of this large hobbyist market, servomotors are readily a vailable in a number of stoc k sizes. Servomotors are used in a few of our robots. Primarily, servomotors are geared dc motors with a positional feedback control that allows the rotor to be positioned accurately . The specifications state that the shaft can be positioned through a minimum of 90° (±45°). In reality we can extend this range closer to 180° (±90°) by adjusting the positional control signal. There are three wire leads to a servomotor . Two leads are for power 15 V and GND. The third lead feeds a position control signal to the motor. The position control signal is a single variable width pulse . The pulse can be varied from 1 to 2 ms . The width of the pulse controls the position of the servomotor shaft. A 1ms pulse rotates the shaft to the extreme counterc lockwise (CCW) position (245°). A 1.5ms pulse places the shaft in a neutral midpoint position (0°). A 2ms pulse rotates the shaft to the extreme CW position ( 145°). The pulse width is sent to the servomotor approximately 50 times per second (50 Hz). Figure 6.27 illustrates the relationship of pulse width to servomotor position. In most of the robots that use servomotors , the servomotor must be posi tioned to its center location before being assembled into the robot. To center the servomotor, we build a simple circuit and PicBasic program. The circuit is shown in Fig. 6.28. The programs for the PicBasic and PicBasic Pro compilers follow: ‘PicBasic program to center servomotorstart:pulsout 0, 150 ‘Send pulse out on rb0pause 18 ‘Delay needed to send pulse at 55 Hzgoto start ‘RepeatTesting the PIC Microcontroller 79 Figure 6.26 Photograph of a servomotor. The following program is for the PicBasic Pro compiler . ‘PicBasic Pro program to center a servomotorstart:pulsout portb.0, 150 ‘Send pulse out on rb0pause 18 ‘Delay needed to send pulse at 55 Hzgoto start ‘RepeatThis centering program and circuit will be referred to in later parts of the book when servomotors are discussed. Parts List LCD serial displayPic Experimenter’s BoardAvailable from Images SI Inc. (see Suppliers at end of book). Microcontroller (16F84) $7.954.0MHz Xtal $2.5080 Chapter SixFigure 6.27 Diagram of pulse widths sent to control servomotor position. Figure 6.28 Schematic of servomotor circuit used to center servomotors.Testing the PIC Microcontroller 81 (2) 22pF capacitors (1) Solderless breadboard RadioShack PN# 276175 (1) 0.1mF capacitor RadioShack PN# 2721069 (2) Red LEDs RadioShack PN# 276208 (2) 470V resistors* RadioShack PN# 2701115 (1) 4.7kV resistor RadioShack PN# 2711126 (1) Voltage regulator (7805) RadioShack PN# 2761770 (1) 9V battery clip RadioShack PN# 270325 Available from RadioShac k, Images SI Inc ., Jameco Electronics , and JDR Microdevices (see Suppliers). *These resistors are also available in 16pin dip package.This page intentionally left blank.7 Chapter Intelligence Programming intelligence into a robot (or computer) is a difficult task and one that has not been very successful to date even when supercomputers are used. This is not to sa y that robots cannot be programmed to perform very useful, detailed, and difficult tasks; they are. Some tasks are impossible for humans to perform quickly and productively. For instance, imagine trying to solder 28 filament wires to a 1/4in square sliver of silicon in 2 s to make an integratedcircuit chip. It’s not very likely that a human would be able to accomplish this task without a machine. But machine task performance, as impressive as it is, isn’t intelligence. Approaches to Building Intelligence There are two schools of thought concerning the creation of intelligence in artificial systems. The first approac h programs an expert system (top down); the second is a neural or beha viorbased system (bottom up). The expert system uses rules to guide the robot in task performance . Behaviorbased programs create an “artificial” behavior in the robot that causes it to reflectively (automatically) perform the task required. Behaviors may be programmed (software) or may be hardwired into the robot. Behaviorbased intelligence doesn’t require a central processor , although such a system ma y have one. Let’s look at a practical programming problem and see how eac h approach differs. Suppose you worked for a company that designed a new robotic vacu um cleaner. The purpose of the robot is to vacuum the floor of a customer’ s home or apartment. Your job is to program the na vigation system. The robot needs to move autonomously throughout the house . How would you go about programming the robot to accomplish na vigation around the home so it could travel in and out of rooms without destroying the place? Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 8384 Chapter SevenLet’s assume you first decide to try an expert na vigation system. This approach uses bruteforce programming and a lot of memory. You might begin by dividing the task of vacuuming the apartment or home into smaller tasks such as vacuuming individual rooms . You begin by programming into the robot’s memory an electronic map (floor plan) of the home or area where the robot needs to vacuum. Then you map out eac h individual room and its contents. The robot must have the ability to measure its movement as it moves as well as compass direction to maintain its location integrity. Once this is accomplished, the robot must have an exact start location on the floor plan. The robot’s movement from the start position is measured and plotted on its internal floor plan map. Problems occur if an object is positioned differently or is out of place, such as a trash receptacle or chair that has been moved. In this situation the real world does not match the robot’s internal map. Similar problems occur if new objects are left on the floor suc h as a bag, toy, or pet. Even so, these obstac les would not present too muc h of a problem for an expert system. To compensate , a secondary collision detection subprogram could be written to detect, map, and go around an obstacle not existing on the internal map. The robot continues to move and vacuum the floor. Keep in mind that as the robot navigates around new obstacles, it’s continually updating its internal map as it tra vels, to maintain its location integrity . These tasks are gobbling up computer time and memory. The robot vacuum accomplished its task. Now suppose you w ant to share this robot or rent it. Now you have a problem. Each new house and every room in the new house would require its own electronic map . Although expert programming does work, it tends to be inflexible and not adaptive tow ard new or innovative situations. Now let’s try the other approach that uses behaviorbased or bottomup programming. Instead of programming internal maps , we program sensor responses and behaviorbased algorithms (feedforward and feedback loops) for sensing and traveling around obstacles and avoiding getting stuck underneath furniture or trapped in corners .Without any internal map we allow the robot to travel and move around the house in a random manner . The idea is that while tra veling in a haphazard manner , it will eventually make its w ay throughout the rooms, cleaning the floor as it goes . Because the robot tra vels randomly, it will take longer for the robot to vacuum the entire floor , and it may miss a spot here and there , but it gets the job done . Since this behaviorbased type of robot vacuum isn’t programmed for a particular house or room, it may be used in any house in any room at any time . While our example is simple, it does illustrate the main differences between expert and beha viorbased (neural) programming . But let’ s look at just one more example before we move on. Expert systems typically have all the answers that the designers believe will be required by the system programmed into the system before it begins. It may store and categorize new information, but based on previously determined categories and existing knowledge . An example of this system could be a roc kIntelligence 85 identification system. The robot examines unknown roc ks based on known characteristics of roc ks, such as color , hardness, scratchability, acid reaction tests, mass, etc. The expert system fails if it inadvertently pic ks up a piece of ice that melts to water during the tests.Well, it fails as long as the designer(s) never anticipated the robot pic king up a piece of ice by mistake and made allowances for it. Neural (behaviorbased) systems are not programmed and are more adaptive, as shown in the previous example. But is a neural system suitable for this task of rock identification? Probably not! There are instances in whic h expert systems are the method of choice. One shouldn’t blindly assume one system is better than the other in all cases . To date, behaviorbased robots are more successful at task accomplishments such as tra veling over unfamiliar and rough terrain than are programmed robots. (Other neuralbased intelligence includes speech recognition, artificial vision, speech generation, complex analysis of stoc k market data, and life insurance policies.) Where’s the Intelligence? Behaviorbased systems at their most basic level are neural reflex actions , so where’s the intelligence in that? However, true behaviorbased systems, when layered on top of one another , generate what appears to be (meaning to us homo sapiens) intelligence actions. This is not a consciousness mind, which is a whole other category of intelligence , but the la yer beha viorbased circuits mimic intelligent actions quite convincingly. Layered Behavioral Responses Let’s layer a few beha vioral responses on top of one another to see how intelligence behavior emerges. This particular robot is a modified “photovore.” It will use a number of standard photoresistors as sensors . Layer 1 is a simple on and off system. It uses a single photoresistor to read ambient light intensity. In darkness the system turns itself off and shuts down all electric power to the robot. When the ambient light increases to a low threshold, the system turns itself on and the robot tra vels forward slowly. Layer 2 is a twophotoresistor sensor . It determines in whic h direction the light intensity is greater. These sensors steer the robot in the direction of the greatest light intensity. Layer 3 is a singlephotoresistor sensor . Under highintensity light it shuts down the robot’s drive system and allows the robot to bathe in strongintensity light. An outsider who didn’t know how this robot was wired would observe the following behavior. At night the robot sleeps. At dawn it begins to travel, looking for a bright light source (food). When it finds a sufficiently bright light source, it stops to feed, recharging its batteries through solar panels.86 Chapter SevenSo our simple photovore robot exhibits three (dare we sa y intelligent?) behaviors—sleep, searching or hunting, and feeding. That’s not bad for a handful of components and some neural glue . BehaviorBased Robotics Behaviorbased programs and robotics are not new concepts . Seminal work has been written and experiments carried out since the 1940s . In the 1940s neural networks and beha viorbased robotics were hardwired electrical components. In the 1940s Dr.W. Grey Walter built two turtlelike mobile robots that exhibit complex beha vior using a few electrical neurons . The beha vior generated was at the time called robotic reflexes. Today this behavior is more accurately described as layered neural architecture. In the 1980s Valentino Braitenberg wrote a book entitled Vehicles— Experiments in Synthetic Psyc hology in which he described complex beha vior emerging from the use of a few artificial neurons . Rodney Brooks, head of MIT’s Artificial Intelligence Laboratory, is a leader in the field of subsumption arc hitecture, which again is beha viorbased and neural. Mark Tilden, creator of the nervous network tec hnology, which again is reflexbased, doesn’t program strategies suc h as w alking into his biomorphic robots. Instead he creates a nervous network whose desired state creates a walking gait. What these scientists ha ve discovered is that neural beha viorbased architecture offers unique advantages over standardbased expert programming . In this book we will build a few behaviorbased robots, using the PIC microcontroller extensively in their construction.Chapter 8 Walter’s Turtle BehaviorBased Robotics Behaviorbased robotics were first built in the 1940s. At that time these robots were described as exhibiting reflexive beha vior. This is identical to the neuralbased approach to implementing intelligence in robots, as outlined in Chap. 7. William Grey Walter—Robotics Pioneer The first pioneer in the bottomup approach to robotics is William Grey Walter. William Grey Walter was born in Kansas City, Missouri, in the year 1910.When he w as 5, his family moved to England. He attended sc hool in the United Kingdom and graduated from King’s College, Cambridge, in 1931. After graduation he began doing basic neurophysiological researc h in hospitals. Early in his career he found interest in the work of the famous Russian psychologist Ivan Pavlov. Do you remember from your high sc hool science classes the famous “Pavlov’s dogs” stimulusresponse experiment? In case you forgot, Pavlov rang a bell just before providing food for dogs . After a while the dogs became conditioned to salivate just by hearing the bell. Another contemporary of Walter, Hans Berger, invented the EEG machine. When Walter visited Berger’ s laboratory, he saw refinements he could make to Berger’s EEG machine. In doing so, the sensitivity of the EEG machine was improved, and new EEG rhythms below 10 Hz could be observed in the human brain. Walter’s studies of the human brain led him to study the neural network structures in the brain. The vast complexities of the biological networks were too overwhelming to map accurately or replicate . Soon he began working with individual neurons and the electrical equivalent of a biological neuron.He wondered what type of behavior could be gathered with using just a few neurons . Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 8788 Chapter EightTo answer this question, in 1948 Walter built a small threewheel mobile robot. The mobile robot measured 12 in high and about 18 in long. What is fascinating about this robot is that by using just two electrical neurons, the robot exhibited interesting and complex beha viors. The first two robots were affectionately named Elmer and Elsie ( electromechanical robot, light sensitive). Walter later renamed the style of robots Mac hina Speculatrix after observing the complex behavior they exhibited. In the early 1940s transistors had not been invented, so the electronic neurons in this robot were constructed by using vacuum tubes. Vacuum tubes consume considerably greater power than semiconductors do , so the original turtle robots were fitted with large rec hargeable batteries. The robot’s reflex or nervous system consisted of two sensors connected to two neurons. One sensor w as a lightsensitive resistor , and the other sensor was a bump switch connected to the robot’s outer housing. The three wheels of the robot are in a triangle configuration. The front wheel had a motorized steering assembly that could rotate a full 360° in one direction. In addition, the front wheel contained a drive motor for propulsion. Since the steering could continually rotate a full 360°, the drive motor’s electric power came through slip rings mounted on the wheel’ s shaft. A photosensitive resistor w as mounted onto the shaft of the front wheel steeringdrive assembly. This ensured that the photosensitive resistor w as always facing in the direction in whic h the robot was moving. Four Modes of Operation While primarily a photovore (lightseeking) type of robot, the robot exhibited four modes of operation. It should be mentioned that the robot’s steering motor and drive motor were usually active during the robot’ s operation. Search. Ambient environment at a low light level or darkness . Robot’s responses, steering motor on full speed, drive motor on 1/2 speed. Move. Found light. Robot’s responses, steering motor off, drive motor full speed. Dazzle. Bright light. Robot’s responses , steering 1/2 speed, drive motor reversed. Touch. Hit obstacle. Robot’s response, steering full speed, reverse drive motor. Observed Behavior In the 1950s Walter wrote two Scientific American articles (“An Imitation of Life,” May 1950; “A Machine That Learns,” August 1951) and a book titled The Living Brain (Norton, New York, 1963). The interaction between the neural system and the environment generated unexpected and complex behaviors. In one experiment Walter built a hutc h, where Elsie could enter and recharge its battery . The hutch was equipped with a small light that wouldWalter’s Turtle 89 draw the robot to it as its batteries ran down.The robot would enter the hutch, and its battery would automatically be recharged. Once the battery recharged, the robot would leave the hutch to search for new light sources. In another experiment he fixed small lamps on each tortoise shell.The robots developed an interaction that to an observer appears as a kind of social beha vior. The robots danced around each other, at times attracted and then repelled, reminding him of a robotic mating ritual or territorial marking beha vior. Building a Walter Tortoise We can imitate most functions in Walter’s famous tortoise . My adaptation of Walter’s tortoise is shown in F ig. 8.1. To fabricate the chassis, we need to do a little metalwork. Working metal is made a lot easier with a few tools suc h as a center punch, hand shears, nibbler, drill, vise, and hammer (see Fig. 8.2). Center punc h: Used to make a dimple in sheet metal to facilitate drilling . Without the dimple, the drill is more likely to “walk” off the drill mark. Hold the tip of the center punc h in the center of the hole you need to drill. Hit the center punch sharply with a hammer to make a small dimple in the material. Shears: Used to cut sheet metal. I would advise purchasing 8to 14in metal shears. Use as a scissors to cut metal. Nibbler: Used to remove (nibble) small bits of metal from sheet and nibble cutouts and square holes in lightgauge sheet metal. Note RadioShack sells an inexpensive nibbler. Figure 8.1 Adaptation of Walter’s turtle robot.90 Chapter Eight Figure 8.2 A few sheet metal tools. Vise: Used to hold metal for drilling and bending. Drill and hammer. Selfexplanatory. A wellstocked hardware store will carry the simple metalworking tools outlined. Most will also carry the lightgauge sheet metal and aluminum bar materials needed to make the c hassis. I built the chassis out of (1/83 1/2in) aluminum rectangle bar and 22to 24gauge stainless steel sheet metal. Stainless steel is harder to work with than cold rolled steel (CRS). And CRS is harder to work with than sheet aluminum. If I were to do this project over , I would use aluminum extensively because it is easier to work with than CRS or stainless steel. Drive and Steering Motors The robot uses servomotors for both the drive and steering . The drive servomotor is a HiT ec HS425BB 51oz torque servomotor (see F ig. 8.3). The HS425BB servomotor is modified for continuous rotation. For steering the robot I used a less expensive HiT ec HS322 42oz torque servomotor (unmodified). Before we go into the robot fabrication, we must first modify the HS425BB servomotor for continuous rotation.Walter’s Turtle 91 Figure 8.3 HS425 servomotor. Modifying the HS425BB Servomotor I chose the HS425BB servomotor because I found it to be the easiest servomotor to modify for continuous rotation. To create a continuous rotation ser vomotor, it is necessary to mechanically disconnect the internal potentiometer from the output gear. First remove the four bac k screws that hold the servomotor together (see Fig. 84). Keep the servomotor horn attac hed to the front of the servomotor . Once the screws are removed, gently pull off the front cover of the servomotor. The output gear will stay attached to the front cover, separating from the shaft of the potentiometer left in the servomotor’s case (see Fig. 8.5). Sometimes the idler gear will fall out. Don’t panic; it’s easy enough to put bac k in position when you reassemble the servomotor. Next remove the plastic clip from the servomotor shaft (see Fig. 8.6).With the plastic clip removed, the shaft of the potentiometer will no longer follow the rotation of the output gear. Align the potentiometer shaft so that the flat sides of the shaft are parallel to the long sides of the servomotor case (see F ig. 8.7). Take off the front cover of the servomotor, and remove the center screw holding the servomotor horn and output gear (see F ig. 8.8). The output gear is92 Chapter Eight Figure 8.4 Removing screws from back of servomotor case. Output Gear Idler Gear Servomotor Horn Plastic Clip Figure 8.5 Inside view of HS425 servomotor.Walter’s Turtle 93 Figure 8.6 Removing plastic clip. Figure 8.7 Top view of servomotor gears with plastic c lip removed. shown in Fig. 8.9. Remove the bearing from the output gear (see Fig. 8.10). The bearing needs to be removed so that you can cut a way the stop tab from the gear. Use a hobby knife or miniature sa w to cut away the stop tab . When you are finished cutting off the tab, check that the cut surfaces are smooth. If not, use a file to smooth out the surfaces . Next remount the bearing onto the gear (see Fig. 8.11). Reassemble the idler and output gears onto the servomotor’s gear train in the case (see Figs. 8.12 and 8.13). Now fit on the servomotor cover , and reattach the cover, using the four screws.94 Chapter Eight Figure 8.8 Removing servomotor horn from front of case . Bearing Stop Tab Output Gear Figure 8.9 Output gear removed from front case.Walter’s Turtle 95Stop Tab Figure 8.10 Stop tab on output that must be removed. Figure 8.11 Stop tab removed and bearing placed bac k on gear.96 Chapter Eight Idler Gear Output Gear Figure 8.12 Output gear fitted back onto servomotor. Figure 8.13 Ready for reassembly of servomotor. The output shaft of the servomotor is now free to rotate continuously . A pulse width of 1 ms sent 50 to 60 times per second (Hz) will cause the servomotor to rotate in one direction. A pulse width of 2 ms will cause the servo motor to turn in the opposite direction. There are two w ays we can stop the servomotor from rotating . The first method is to simply stop sending pulses to the servomotor. The second method is a little tric kier. A pulse width of approximately 1.5 ms will stop the servo Walter’s Turtle 97 motor. The exact pulse width for eac h servomotor must be determined experimentally. The exact pulse width required is based upon the position of the static potentiometer shaft inside the servomotor . If you followed the directions provided, it should be about 1.5 ms . To find the exact pulse width to stop the servomotor, you have two options. The first is to keep manually adjusting the pulse width until you find the correct pulse width. As you approach the pulse width needed to stop the servomotor, you will notice that the rotational speed of the servomotor will slow down. You can use this as a feature to create a speed control, if you wish. The second option is to look at the servomotor circuit described in Chap. 14 (see Fig. 14.11). This simple circuit allows you to quickly find the correct pulse width. Sheet Metal Fabrication There are three pieces of sheet metal one needs to fabricate . The U bracket, shown in Fig. 8.14, holds the front wheel and drive servomotor. The U brac ket may be fabricated from 22gauge 1.253 5in aluminum sheet metal. I would recommend purc hasing the U brac ket (see P arts List) because the cutting required for this fabrication is extensive and precise . The U brac ket mounts the drive servomotor (see F ig. 8.15). In addition, on the top of the U brac ket are holes for mounting a servomotor horn, which is used to connect the steering servomotor. Figure 8.16 is a diagram of the base with a cutout for the 42oz servomotor . The base measures 3 in 3 5.5 in. The base will hold the power supply and the electronics. Follow the servomotor diagram in removing metal from the base . First drill the four ( 1/8in) holes for mounting the servomotor . Next use the same drill bit to drill holes along the inside perimeter of the servomotor cutout. Removing metal in this way is a little easier than trying to saw or nibble it away. When you ha ve drilled as many holes as possible , use the metal nibbler to cut the material between the holes to finish removing this material. Then continue to nibble away at the sides of the cutout until you have the rectangle shape needed. Before you mount the servomotor , file the edges of the hole smooth. Finish the base by drilling the other holes outlined in the dra wing. The rear axle bracket is shown in F ig. 8.17; it is made from 1/83 1/23 10in aluminum bar. Drill the four 1/8in holes in the aluminum before bending it into shape. For the rear axle I used the wire from a metal coat hanger . Mount the rear axle and wheels to the robot base, using two 632 machine screws and nuts. To continue, we need to mount the front drive wheel to the servomotor . The drive wheel has a diameter of 23/4 in and is 1/8 in thick (see Fig. 8.18). The holes are drilled in the wheel to accept a standard HiT ec servomotor horn (see F ig. 8.19). The horn is secured to the wheel using four no . 2 3 1/4in sheet metal screws (see Fig. 8.20). Before you attach the servomotor to the U bracket, secure a servomotor horn to the top of the U brac ket, using the predrilled holes (see F ig. 8.21).Fits HS-322, HS-325, HS-425 servomotors 5.2 5.80 1.25 2.37 3.57 .625 Hole size 5/32 2.97 Bend 90° Bend 90° Material 1.25 x 5.80 x .050 Aluminum 6061 7/32 dia. hole Bracket bent at 90° .224 C/L 1.53 .37 C/L 2.14 C/L Rectangle cutout .83 x 1.62 .29 dia. .6 .288 .39 .8 .91 .14 .31 .625 .94 1.11 Four small shaded holes 1/8 dia. All dimensions in inches 98 Chapter Eight Figure 8.14 Drawing of U bracket for mounting the drive servomotor.Walter’s Turtle 99 Figure 8.15 U bracket with drive servomotor attached. The front of the mounting ears, both top and bottom, on the servomotor has small tabs (see Fig. 8.22). Cut and file away these tabs so that the servomotor can be mounted flush against the brac ket (see Fig. 8.23). Next mount the servomotor to the U brac ket, using 632 mac hine screws and nuts . Attach the wheel/horn assembly to the servomotor (see F igs. 8.24 and 8.25). Put this assembly to the side while we work on other components . Shell The original tortoises used a transparent plastic shell. The shell was connected to a bump switc h that caused the robot to go into “avoid” mode when activated. I looked at, tried, and rejected a number of different shells. Finally I was left with no choice other than to fabricate my own shell. Rather than fabricate an entire shell, I made a bumper that encompasses the robot. The bumper is fabricated from 1/83 1/23 32in aluminum bar (see Fig. 8.26). The aluminum bar is marked at the center . Each bend required in the bumper is also marked in pencil. The material is placed in a vise at eac h pencil mark and bent to the angle required.The two ends of the aluminum bar end up at the center bac k of the bumper. These two ends are joined together using a 1/83 1/23 1inlong piece of aluminum bar . A 1/8in hole is drilled on each end of the aluminum bar . Matching holes are drilled in the ends of the42-oz servo 1/2 Ø wire pass-through hole 1/8 -in holes to match rear axle bracket Sheet metal 3 in x 5.5 in Front 100 Chapter Eight Figure 8.16 Robot base showing cutout for 42oz servomotor and holes for rear axle bracket. bumper. The bar is secured to the bumper using two 540 machine screws and nuts (see Fig. 8.27). The upper bracket used to connect the bumper to the robot is identical to the front end of the bumper (see F ig. 8.28). The upper bracket is made from 1/83 1/23 14.5in aluminum bar . As with the bumper , the center of the bar is marked, and each bend required is also marked in pencil. The material is bent in a vise the same way as the bumper. Finding the Center of Gravity It is important to find the center of gravity line of the bumper, because this will mark the optimum location where the upper bracket should be attached. Rest the bumper on a length of aluminum bar . Move the bumper bac k and forth until it balances evenly on the aluminum bar. Mark the centerline positions on each side of the bumper. Drill a 1/8in hole on each side. Drill matching holes onWalter’s Turtle 101 3/4 in 23/4 in1/8 -in hole for axle 1/8 -in holes 1/8 -in hole for axle 125° 23/4 in 35/8 in Wheel Wheel size Axle Height 2 in 10 in Figure 8.17 Rear axle bracket detail. the ends of the upper brac ket. Then secure the upper brac ket to the bumper using 540 machine screws and nuts. Attaching Bumper to Robot Base The bumper is attached to the robot body by the upper bracket. Drill three 1/8in holes in the top of the upper bracket. One 1/8in hole is in the center, and the two other holes are 11/8 in away from the center hole (see Fig. 8.29). Three matching holes are drilled in the robot base behind the servomotor . The holes should be placed so that the bumper (once secured to the base) has adequate clearance (1/8 to 1/4 in) from the back wheels. The matching center hole on the base must be offset by moving the drilled hole forw ard on the base by about 1/4 in.102 Chapter Eight .575 .575 .385 .385 Center hole .300 diameter All other holes 1/8 diameter Material 1/8 -thick hardwood 2.75 diameter 1/8 Side view All dimensions in inches Figure 8.18 Drawing of drive wheel. Figure 8.19 Servomotor drive wheel with holes for mounting servomotor horn.Walter’s Turtle 103 Figure 8.20 Drive wheel with servomotor horn attac hed. Figure 8.21 U bracket before mounting of drive servomotor. The bracket is secured to the base using two 1inlong 632 machine screws, four 632 nuts, and two 1inlong 2lb compression springs, with a 1/8in center diameter (see F ig. 8.30). The tension and resiliency of the bumper can be adjusted by tightening or releasing the upper 632 mac hine screw nuts. Once assembled, the bumper will tilt bac k and c lose the bumper switc h when the robot (bumper) encounters (pushes against) an obstac le.Tab Tab Filed Away 104 Chapter Eight Figure 8.22 Tab on servomotor case that needs to be filed off . Figure 8.23 Tab files off servomotor case. Bumper Switch The bumper switch makes use of the center holes . Looking back at F ig. 8.30, we see the center hole is fitted with a 632 mac hine screw held on by a standard (zincplated) nut, followed by a brass nut. The brass nut has a wire soldered to it. The purpose of this little assembly is just to attac h a wire to the bracketbumper assembly. Brass nuts are used because it is possible to solder wires to brass to make electrical connections . This is in contrast to the standard zincplated steel nuts that are very difficult (impossible) to solder .Walter’s Turtle 105 Figure 8.24 Attaching drive servomotor to U brac ket by using plastic screws and nuts. Figure 8.25 Another view of drive servomotor and U brac ket.106 Chapter Eight 91/2 in 43/4 in 43/4 in 41/2 in 41/2 in 31/2 in 32 in Figure 8.26 Top dimensional view of bumper fabricated from 1/8in 3 1/2in 3 32in aluminum bar. Aluminum bumper 5-40 nuts 1-in-long aluminum bar 5-40 machine screws Figure 8.27 Cutaway closeup of aluminum bracket used to secure the open ends of the bumpers. The second half of the tile switch is comprised of a 1in 632 plastic machine screw and three 632 machine screw nuts. One nut must be brass with a wire soldered to it (see Fig. 8.31). Figures 8.32 and 8.33 are closeup photographs of the finished bumper switch. The assembly is adjusted so that the brass nut on the top of the 632 mac hine screw lies just underneath the upper aluminum bracket without touc hing. When the upper brac ket tilts forw ard, contact is made between the aluminum bracket and brass nut, which is read as a switch closure.Walter’s Turtle 107 141/2 in 1 in 1 in 41/2 in 41/2 in 31/2 in Figure 8.28 Side dimensional view of upper brac ket fabricated from 1/8in 3 1/2in 3 141/2in aluminum bar. 11/8 in 11/8 in 1/8 -in hole Figure 8.29 Side dimensional view for hole placement in top of the upper brac ket. 6-32 machine screw 6-32 nut 6-32 nut 6-32 nut 6-32 brass nut Soldered wire 1-in-long compression spring 1-in-long compression spring 6-32 1/2machine screw Figure 8.30 Side view of upper bracket detailing the mounting of the upper bracket to the robot base using mac hine screws and compression springs . Also details bracket half of the bumper switc h. Mounting the Steering Servomotor If you haven’t done so, mount the steering servomotor to the robot base, using four 632 plastic machine screws and nuts. Before you attach the U bracket to the steering servomotor, make sure the steering servomotor spindle is in its center position. This will ensure that the robot will steer forward right and left108 Chapter Eight 6-32 brass nut 6-32 nuts Upper bracket Wire 6-32 plastic machine screw Base Robot base Figure 8.31 Side dimensional detail (robot base side of the bump switc h) of plastic screw with top brass nut. Figure 8.32 Closeup photograph detailing bump switch and spring mounting of upper brac ket. properly. The following short program will place a servomotor in its center position: start:pulsout portb.1, 150pause 18goto startThe output pulse signal for the servomotor is taken as pin RB1. Once the servomotor is in its center position,attach the U bracket to the servomotor so that the drive wheel is pointing forw ard.PhotoresistorThe CdS photoresistors (see Fig. 8.34) used in this robot have a dark resistance of about 100 kV and a light resistance of 10 k V. The CdS photoresistors ha ve large variances in resistance between cells. It is useful to use a pair of CdS cells for this robot that matches, as best as one can matc h them, in resistance. Since the resistance value of the CdS cells can vary so greatly , it’s a good idea to buy a few more than you need and measure the resistances , to find a pair whose resistances are c lose. There are a few w ays you can measure the resistance. The simplest method to use a voltohmmeter, set to ohms. Keep the light intensity the same as you measure the resistance . Choose two CdS cells that are closely matched within the group of CdS cells you ha ve. The second method involves building a simple PIC16F84 circuit connected to an LCD displa y. The advantage of this circuit is that you can see the response of the CdS cells under varying light conditions . In addition, you can see the difference in resistance between the CdS cells when they are held under the same illumination. This numeric difference of the CdS cells under exact lighting is used as a fudge factor in the final turtle program. If you just test the CdS cells with just an ohmmeter, you will end up using a larger fudge factor for the robot to operate properly. The schematic for testing the CdS cells is shown in F ig. 8.35. The circuit, built on a PIC Experimenter’ s Board, is shown in F ig. 8.36. The PicBasic Pro testing program follows: ‘CdS cell test ‘PicBasic Pro program ‘Serial communication 1200 baud true ‘Serial information sent out on port b line 0 Walter’s Turtle 109 Figure 8.33 Closeup photograph detailing bump switch.110 Chapter Eight Figure 8.34 CdS photoresistor cell. LCD Display V1 100KW V2 100KW CdS Cell CdS Cell C2 .1μF 50V C3 .1μF 50V SW4 C1 .1μF R1 4.7KW U1 +5V X1 4MHz 416 15 PIC 16F84 5 VSS VDD 17 181 2 3 6 7 8 910 11 12 13 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA014 MCLR' OSC1 OSC2 Serial Line +5V Gnd Figure 8.35 Electrical schematic for testing and calibrating CdS cells . ‘Read CdS cell #1 on port b line 1 ‘Read CdS cell #2 on port b line 7 v1 var byte ‘Variable v1 holds CdS #1 information v2 var byte ‘Variable v2 holds CdS #2Walter’s Turtle 111 Figure 8.36 Test circuit built on PIC Experimenter’s Board. information pause 1000 ‘Allow time for LCD display main: pot portb.1,255,v1 ‘Read resistance of CdS #1 photocell pot portb.7,255,v2 ‘Read resistance of CdS #2 photocell ‘Display information serout portb.0,1,[$fe,$01] ‘Clear the screen pause 25 serout portb.0,1,[“CdS 1 = ”] serout portb.0,1,[#v1] serout portb.0,1,[$fe,$C0] ‘Move to line 2 pause 5 serout portb.0,1,[“CdS 2 = ”] serout portb.0,1,[#v2] pause 100 goto main Notice in Fig. 8.36 that CdS cell 1 is reading 37 and CdS cell 2 is reading 46 under identical lighting . Keep in mind that this is a c losely matched pair of CdS cells.We can use a fudge factor of ±15 points. This means that as long as the readings between cells vary from eac h other by ±15 points , the microcontroller will consider them numerically equal.112 Chapter Eight Trimming the Sensor Array If you are using the Experimenter’ s Board, you can trim and matc h the CdS cells to one another . Doing so allows you to reduce the fudge factor and produces a crisper response from the robot. Typically one CdS cell resistance will be lower than that of the other CdS cell. To the lowerresistance CdS cell add a 1k V (or 4.7kV) trimmer potentiometer in series (see Fig. 8.37). Adjust the potentiometer (trim) resistance until the outputs shown on the LCD displa y equal eac h other. Trim the CdS cell under the same lighting conditions in whic h the robot will function. The reason for this is that when the light intensity varies from that nominal point to whic h you’ve trimmed the CdS cell, the responses of the individual CdS cells to c hanges in light intensity also vary from one another and then are not as c losely matched. Once you have a pair of CdS cells to use , they need to be attac hed to the robot. I soldered the CdS cells and capacitors to a small piece of perforated board (see Fig. 8.38). Figure 8.38 shows both the front and bac k of the sensor array. The opposite side of the servomotor brac ket that holds the continuous rotation servomotor is perfect for mounting the photoresistor. I used a small piece of transparent plastic, 1/2 in wide 3 6 in long 3 1/16 in thick (12.5 mm 3 152 mm 3 1.5 mm thick) to create an L brac ket on which to mount the photoresistors (see Fig. 8.39). A 1/8in hole is drilled 1/2 in up from one end (see Fig. 8.37). The plastic is then gently heated about 2 1/2 in up from the end (see bend point). When the plastic softens, bend it to a 90° angle and hold it in position until the plastic hardens again. LCD Display V1 100KW V2 100KW CdS Cell CdS Cell C2 .1μF 50V C3 .1μF 50V SW4 C1 .1μF R1 4.7KW U1 +5V X1 4MHz 416 15 PIC 16F84 5 VSS VDD 17 181 2 3 6 7 8 910 11 12 13 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA014 MCLR' OSC1 OSC2 Serial Line +5V Gnd 1KW V3 Figure 8.37 Electrical schematic of testing circuit with potentiometer trimmer.BACK FRONT Walter’s Turtle 113 Figure 8.38 Front and bac k mounting of CdS cells and capacitors to perforated board. Bend point 90° 31/2 in 6 in 1/2 in 21/2 in 31/2 in Material: Plastic Size 1/2 in x 6 in x 3/32 in Figure 8.39 Fabrication drawing for plastic bracket for CdS cells.114 Chapter Eight Figure 8.40 CdS sensor array attached to plastic bracket. Next I used hot glue to secure the CdS cells to the bac k of the plastic L (see Fig. 8.40). Then I mounted an opaque vane on the front surface of the plastic in between the photoresistors (see Fig. 8.41). The opaque vane is made from a small piece of conductive foam I had lying around. I simply hotglue one edge to the plastic. Using the opaque vane and the two CdS photosensors in this configuration alleviates much of the computation needed to track a light source. The operation of the sensor array is shown in Fig. 8.42. When both sensors are equally illuminated, their respective resistances are approximately the same. As long as each sensor is within ±10 points of the other, the PIC program will see them as equal and won’t move the servomotor (steering). When the sensor array is not properly aimed at the light source, the vane’s shadow falls on one of the CdS cells. This pushes the resistance beyond the ±10point range. The PIC microcontroller activates the steering servomotor to bring both sensors bac k under even illumination. In doing so, this steers the robot straight to the light source . If the sensors detect too great a light intensity, the robot will go into avoid mode. Mounting the photoresistor arra y on the drive wheel assembly keeps the sensors pointing in the same direction as the drive wheel (see F ig. 8.43). This replicates the function of the original tortoise robots . The array is secured to the U bracket by using a small plastic screw and wing nut. Schematic The schematic for the robot is shown in F ig. 8.44. Intelligence for the robot is provided by a single PIC 16F84 microcontroller . The forw ard servomotor isWalter’s Turtle 115 SIDE VIEW TOP VIEW CdS Cell CdS Cell 1/2 in Vane Plastic "L" Perf. Board 21/2 in Figure 8.41 Drawing showing CdS cells attached to bracket with vane. connected to RB7, and the steering servomotor control signal is provided by RB6. Sensor readings of the CdS cell are read off pins RB2 and RB3. The bumper switch is read off pin RA0. There is nothing critical about the circuit; it may be hardwired on a prototyping board. I chose a simpler route . Images SI Inc . sells a four servomotor controller board. This board has all the connections needed for the sensors and servomotors. My connections to the PC board are shown in Fig. 8.45. A picture of the finished circuit is shown in Fig. 8.46. Notice in the picture I used terminal blocks to connect the sensor array and bumper switch. Program Upon power up, the drive motor is off, and the microcontroller begins scanning for the brightest light source, using the servomotor. If a light source is too bright,the robot jumps into avoid mode. In avoid mode the robot backs away from the light source by reversing the drive motor while steering the drive wheel left or right. If the light isn’t so bright as to activate116 Chapter Eight A B A B Light source A cell in shadow; tracker rotates to right. B cell in shadow; tracker rotates to left. Equal illumination; no movement. SIDE VIEW Figure 8.42 Operation of sensor array for targeting light source. the avoid mode, the robot steers in the direction of the light and activates the drive wheel forward. If the bumper switc h is activated, the robot assumes it has hit an obstac le and so goes into a void mode. The robot uses a void mode for too bright a light and collisions. If the tilt switch is not activated (no collision),then the program jumps to the beginning and the process continues scanning and moving to the brightest light source. The program is written for the PicBasic Pro compiler that is programmed into a PIC 16F84. The program should be able to be compiled and run with few modifications on the PicBasic version. Ingroup variances in CdS sensors, drive motors, robot structure, and the like can be adjusted for or modified in the program. ‘Turtle program‘PicBasic Pro program‘Read CdS cell #1 on port b line 1‘Read CdS cell #2 on port b line 7v1 var byte ‘Variable v1 holds CdS #1 informationv2 var byte ‘Variable v2 holds CdS #2 informationv3 var byte ‘Variable for calculation s1 var byte ‘Variable s1 holds servomotor #1 pulse width info s2 var word ‘Variable for random function rv var byte ‘Variable rv holds the range value s1 = 150 ‘Initialize steering servomotor facing forward rv = 10 ‘Adjust as needed for smooth operation ct var byte ‘Counter ‘Drive servomotor ** continuous rotation information ‘Connected to pin portb.7 ** variable pulse width numbers ‘157 forward * 165 slow forward ‘167 stop ‘169 slow backward * 177 backward start: pot portb.2,255,v1 ‘Read resistance of CdS #1 photocell pot portb.3,255,v2 ‘Read resistance of CdS #2 photocell ‘Check bumper switch “Did I hit something?” if porta.0 = 0 then avoid ‘Hit obstacle go into avoid mode ‘Is it sleepy time? if v1 <= 230 then skp ‘Is it dark enough to sleep? Walter’s Turtle 117 Figure 8.43 Attaching sensory array to drive servomotor’s U bracket.118 Chapter Eight Figure 8.44 Schematic of robot. if v2 > 230 then slp ‘Is it too bright to see?skp: if v1 >= 12 then skip2 if v2 < 12 then avoid ‘Which way do I go?skip2: if v1 = v2 then straight if v1 > v2 then greater if v1 < v2 then lesser straight: pulsout portb.6, s1 pulsout portb.7, 157 goto startgreater:v3 = v1 v2 if v3 > rv then right goto straight lesser:V1 100KW V2 100KW CdS Cell CdS Cell C2 .1μF 50V C3 .1μF 50V SW4 C1 .1μF R1 4.7KW U1 +5V X1 4MHz 416 15 PIC 16F84 5 VSS VDD 17 181 236 7 8 910 11 12 13 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA014 MCLR' OSC1 OSC2 +5V +5V +5V R2 10KW SW3 Bumper switch Servo Motor 1 Servo Motor 2 ‘Yes‘No sleep––keep moving‘Is it too bright to live?‘Yes‘Not so bright––should I steer?‘Light is equal go straight‘Check light intensity to turn right‘Check light intensity to turn left‘Go forward in the direction you’re facing‘Don’t move steering‘Go forward‘Check numerical difference between CdS cells‘If more than rv turn right‘If not go straightWalter’s Turtle 119 1 2 3 4 Images SI Inc., NY Top Bottom 9V Battery Servomotor Connector S + – Servomotor Signal +5V Ground S + – S + + – – S + – S + – 470 10K 10K 4.7K Ser Out Gnd U1 16F 84 Reset Gnd Gnd Gnd 1 2 3 4 Servomotor Controller To Bumper Switch To Sensor Array On Off C1 C2 C1 DC Power Jack PJ-102B + – Servomotor #1 Forward Servomotor #2 Turn Figure 8.45 Using an existing PCB board for building robot’ s electronics. v3 = v2 v1 ‘Check numerical difference between CdS cells if v3 > rv then left ‘If more than rv turn left goto straight ‘If not go straight right: ‘Turn right s1 = s1 + 1 ‘Increment variable s1 to turn right if s1 > 225 then s1 = 225 ‘Limit s1 to 225 pulsout portb.6, s1 ‘Move steering servomotor pulsout portb.7, 165 ‘Go forward slowly goto start left: ‘Turn left120 Chapter Eight Figure 8.46 Closeup of electric circuit board. s1 = s1 1 if s1 < 65 then s1 = 65 pulsout portb.6, s1 pulsout portb.7, 165 goto startslp: pulsout portb.6, s1 pulsout portb.7, 167 goto startavoid: random s2 s1 = s2 /256 if s1 < 65 then s1 = 65 if s1 > 225 then s1 = 225 for ct = 1 to 125 pulsout portb.6, s1 pulsout portb.7, 177 pause 18 next ct s1 = 150 ‘Decrement variable s1 to turn left‘Limit s1 to 65‘Move steering servomotor‘Go forward slowly‘Go asleep‘Don’t move steering‘Stop drive servomotor‘Avoid mode, send‘Randomize s2‘Reduce range of s1 to 1 to 255‘Set lower limit‘Set upper limit‘Start counter‘Steer (turn) in a random direction‘Reverse drive motor (slow)‘Pause to send instructions at 50 Hz‘Loop‘Steer back to centerWalter’s Turtle 121 goto startAdding Sleep Mode I added a sleep mode for occasions when the ambient light is very low . The robot moves forward when both CdS sensors receive approximately the same light intensity. The robot steers right or left when one CdS cell receives more light than the other . If eac h CdS cell receives too muc h light or the bump switch is activated, the robot goes into avoid mode. Power A 9V battery on the PC board supplies adequate electrical power for the robot for a short time. Although I used this power supply for testing robot function, you will need a stronger power supply for extended use . The PCB board has a dc voltage socket where an external power supply can be connected. The finished robot is shown in F igs. 8.47 and 8.48. Behavior This robot exhibits the following beha vior. In ambient light, no bright light source, the robot tra vels in a straight line (or circ le depending upon the last light source target). If the ambient light is too bright, it jerks backward.With a mediocre light source, it will aim and travel toward the light. The program can be developed further to explore more interesting and exotic behaviors. Before we do so, let’s first look at how the standard program functions. Fudge FactorThe variable RV (range value) is the fudge factor. At the beginning of the program the variable RV is assigned a value of 10.In my prototype I actually used an RV of 2 because I had matc hed the resistance values of CdS cells , as discussed earlier. Tolerance between the two CdS photoresistors ma y be increased or decreased by modifying the numerical value of this variable . You may need to adjust this variable according to how closely the resistance values of your CdS cells match. Light Intensity The program continually c hecks the light intensity received (resistance) by each CdS sensor and then makes a decision based on those readings. The maximum reading from the sensor is 255 (total darkness). If the room gets dark enough to generate a value of 230 in eac h CdS cell, then the robot goes into sleep mode. ‘Is it sleepy time?122 Chapter Eight Figure 8.48 Side view of turtle robot. Figure 8.47 Front view of turtle robot.Walter’s Turtle 123 if v1 <= 230 then skp ‘Is it dark enough to sleep?if v2 > 230 then slp ‘YesThe opposite of sleep mode is the “too bright to live.” If the light becomes too intense, this triggers the avoid mode. ‘Is it too bright to see?skp: ‘No sleep—keep moving if v1 >= 12 then skip2 ‘Is it too bright to live? if v2 < 12 then avoid ‘Yes Increasing the numerical value, in this case 12, decreases the light intensity that puts the robot into a void mode. Decreasing the numerical value increases the light intensity needed to throw the robot into avoid mode. In most cases you will want to decrease this number. However, I would advise you not to go below a numerical value of 9, because even at full light saturation of the CdS cell, its resistance never drops to zero. And in my light saturation tests the sensor never yielded a value less than 5. In this robot, intense light pushes the robot into a void mode. If this were a true photovore robot, high light intensity would put it into a feeding mode . HandednessIn addition, one could create handedness in the robot (rightor lefthanded) by modifying either greater or lesser subroutines , not both. This will create a robot that is more likely to turn in one direction than in the other . greater:v3 = v1 v2 ‘Check numerical difference between CdS cellsif v3 > rv then right ‘If more than rv turn rightgoto straight ‘If not go straightFor instance, if RV 5 10, we can substitute the value 7 like this if v3 > 7 then rightThen in the lesser subroutine the RV is kept the same . The result of this manipulation is that we would create a robot that is more likely to turn to the right. This robot offers opportunities to the robotists and experimenters for continued experimentation and development in both hardw are and software. Parts List 12in 3 12in sheet metal sheet of 22 or 24 gauge (1) Aluminum bar 1/8 in 3 1/2 in 3 32 in long (1) Aluminum bar 1/8 in 3 1/2 in 3 141/2 in long (1) Aluminum bar 1/8 in 3 1/2 in 3 2 in long124 Chapter Eight (1) 42oztorque hobby servomotor (HS322) (1) Hobby servomotor (HS425) (1) 23/4indiameter drive wheel (2) CdS photocells, 100 kV dark, 10 kV light (1) 10kV, 1/4W resistor (1) 4.7kV, 1/4W resistor (2) 22pF caps (1) 4MHz ceramic resonator or Xtal (1)(IC1) PIC microcontroller (16F8404)(1) U bracket for drive servomotor Miscellaneous needs include perforated board, 1/16inthick transparent plastic, 540 machine screw and nuts , plastic 632 3 1in machine screw, 632 brass nuts, 1inlong compression springs (2 lb). Aluminum bars, machine screws, tubing, and compression springs are a vailable in most wellstocked hardware stores. Servomotors may be purchased at hobby shops or electronics distributors . Electronic components may be purchased from RadioShack, Images SI Inc., Jameco Electronics, JDR Microdevices (see Suppliers at end of book). PC board, servomotor drive wheel, and U bracket for drive servomotor may be purchased from Images SI Inc.9 Chapter Braitenberg Vehicles In 1984 Valentino Braitenberg published a book titled Vehicles—Experiments in Synthetic Psychology. In his book Valentino describes a number of wondrous vehicles that exhibit interesting beha viors based on the use of a few electronic neurons. Similar in concept to Walter’s seminal neural work with his robot tortoises , Valentino’s vehicle behavior is more straightforward, making it somewhat easier to follow both theoretically and logically. This also makes it easier to implement his ideas into real designs for robots . In this chapter we will build a few Braitenbergtype vehic les. At the heart of Braitenberg vehic les is his description of a basic vehic le, which is a sensor connected to a motor . Braitenberg continues to explain the relationship between the sensor and motor. The relationship is essentially the connection between the sensor and motor, and this connection ought to be considered as a neuron.With the connection configured as a neuron,the structure is shown in Fig. 9.1. Instead of a vehicle we will describe the structure diagram as a small neural network. At the front end of the network we find a sensor , followed by the neuron and finally the output motor . The sensor detects the intensity of light and outputs a proportional signal to the motor . Highintensity light produces high rpm’s (revolutions per minute) from the motor. Lowintensity light produces slow rpm’s. Consider the sensor portion as modular and interchangeable. Other sensors can be plugged in and incorporated to detect any number of environmental variables, for example, heat, pressure, sound, vibration, magnetic fields (compass), electrical fields, radioactivity, and gases (toxic or otherwise). In addition, the motor, like the sensor, represents a singular example of an output module. Other output modules could include a second neuron (or neural layer), electric circuit, on/off switch, light source, etc. Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 125126 Chapter Nine Figure 9.1 Basic neuron setup, sensor input, neuron, and motor output. The neuron’s input is the output of the sensor , and the neuron’ s output activates a motor in relationship to its input. The input/output “relationship” of the neuron can be made to be one of many different mathematical functions. The relationship ma y also be called connection strength or connection function when you are reading the neural network literature . The relationship is one of the most important variables we can modify when programming our robot. Neural I/O Relationships When the neuron is stimulated, it generates an output. As stated, there are a number of mathematical functions that can exist inside the neuron. These functions act upon the neuron’ s input (sensor output) and pass through the results to the neuron’s output. Let’s examine a few of them. Positive proportional. As input from the sensor increases , activation (rpm’s) of the motor increases in proportion; see Fig. 9.2.Negative proportional. As input from the sensor increases, activation (rpm’s)of the motor decreases in proportion (see Fig. 9.3).Digital. As input from the sensor output exceeds a predetermined (programmed) threshold (that ma y be positive or negative), the motor is activated (see Fig. 9.4). Gaussian. As input from the sensor increases , output passes through a gaussian function for motor activation (see Fig. 9.5). Essentially the neuron ma y incorporate any mathematical function. It would perform this function on the sensory input to generate an appropriate output. I have provided an example of only a few of the more common functions available. Vehicles Using the basic neural setup , we can construct a few simple vehic les that exhibit interesting behaviors. Figure 9.6 illustrates two vehicles labeled A and B. Both vehicles use the positive proportional neural setup with a light intensity sensor.Vehicle A, if both sensors are evenly illuminated by a light source, will speed up and, if possible, run into the light source. However, if the light source is off to one side, the sensor on the side of the light source will speed a little faster than the sensor/motor on other side . This will cause the vehic le to veer away from the light source (see F ig. 9.7). Vehicle B, if both sensors are evenly illuminated by a light source , will speed up and, if possible , run into the light source (same as vehic le A). If the light source is off to one side, vehicle B will turn toward the light source (see Fig. 9.7). Braitenberg Vehicles 127 Figure 9.2 Graph of positive proportional transfer function. As sensor output increases , motor output increases. Figure 9.3 Graph of negative proportional transfer function. As sensor output increases , motor output decreases. Figure 9.4 Graph of digital transfer function. As sensor output increases, output remains unchanged until threshold is reached, then output switches full on.128 Chapter Nine Figure 9.5 Graph of gaussian function. As sensor output increases, output follows a gaussian curve. Figure 9.6 Wiring of two Braitenberg vehicles labeled A and B. Negative proportional neural setups would show the opposite beha vior. Building Vehicles It’s time to put the theory to the test and see if it works . Let’s assemble the materials needed to build a vehicle. The photovore’s basic operating procedure is like Walter’s robot. It tracks and follows a light source . The base of the vehic le is a sheet of aluminum 8 in long by 4 in wide by 1/8 in thick. We will use two gearbox motors for propulsion and steering and one multidirectional front wheel. We will try a new construction method with this robot.Instead of securing the gearbox motors with mac hine screws and nuts , we will use 3M’ s industrial brand doublesided tape. This doublesided tape, once cured, is as strong as pop rivets. I tried to separate a sample provided by 3M.It consisted of two flat pieces of metal secured with the tape . Even when I used pliers , it was impossible. 3M states that the tape requires 24 h to reach full strength.You may not achieve the fullstrength capability of the tape unless you follow the 3M procedure .Braitenberg Vehicles 129 Figure 9.7 Function of A and B Braitenberg vehicles. The gearbox motor is a 918D type (see F ig. 9.8). The gearbox motor at the top of the picture has an orange cowl that is covering the gears. Notice the flat mounting bracket that is perfect for securing to the vehic le base. The doublesided tape is cut lengthwise to fit the base of bracket to the gearbox motor. The exposed side of the tape is immediately secured to the gearbox motor brac ket. Then the motor is positioned on the bottom of the vehic le base, the protective covering of the tape is removed, and the gearbox motor is firmly placed onto the bottom of the vehicle base (see Fig. 9.9). The second gearbox motor is secured to the other side in a similar manner . Back wheels The shaft diameter of the gearbox motor is a little too small to make a good friction fit to the rubber wheel. To beef up the diameter, cut a small 1to 1.5in length of the 3mm tubing; see Parts List. Place the tubing over the gearbox motor shaft, and collapse the tubing onto the shaft, using pliers. There is a small cutaway on the gearbox motor shaft (see F ig. 9.10). If you can collapse the tubing into this cutaway, you will create a strong fit between the shaft and the tubing that will not pull off easily (see F ig. 9.11). The tubing adds to the diameter of the shaft and will make a good friction fit with the rubber wheels (see F ig. 9.12). Simply push the center holes of the wheels onto the tubing/shaft, and you are finished.130 Chapter Nine Figure 9.8 A 918D 100:1 Gearbox motor. Figure 9.9 3M doublesided tape is used to secure gearbox motor to base of vehicle.Braitenberg Vehicles 131 Figure 9.10 Gearbox motor showing cutaway on output shaft. Figure 9.11 A 11/2in length of 3mmdiameter tubing attached to gearbox motor shaft. Front wheels Steering is accomplished by turning on or off the gearbox motors . For instance, turning on the right while the left gearbox motor is off will turn the vehicle to the left, and vice versa. In similar vehicles many times the robotists will forgo front wheels entirely and use a skid instead. This allows the vehic le to turn without concern about the front wheels pivoting and turning in the proper direction The multidirectional wheel accomplishes much the same thing as a skid,but does so with less resistance . Figure 9.13 shows the multidirectional wheel. It is constructed using rollers around its circumference that allow the wheel to rotate forward and move sideways without turning.132 Chapter Nine Figure 9.12 Rubber wheel used to friction fit onto gearbox motor shaft. The multidirectional wheel is attac hed using a basic Ushaped brac ket (see Fig. 9.14). The bracket is secured to the front of the vehic le base using the 3M doublesided tape. The multidirectional wheel is secured inside the U brac ket using a small 2.25in piece of 1/420 threaded rod and two mac hine screw nuts (see Fig. 9.15). With the motors and the multidirectional wheel mounted, we are ready for the electronics . Figure 9.16 shows the underside of the Braitenberg vehicle at this point. I drilled a 1/4in hole in the aluminum plate to allows wires from the gearbox motors underneath the robot to be brought topside. The schematic for the electronic circuit is shown in F ig. 9.17. I built the circuit on two small solderless breadboards. You can do the same or hardwire the components to a PC board. The circuit is pretty straightforw ard. The gearbox motors require a power supply of 1.5 to 3.0 V. Rather than place another voltage regulator into the circuit, I wired three silicon diodes in series off the 5V dc power. The voltage drop across each diode is approximately 0.7 V. Across the three series diodes (0.7 3 3 5 2.1 V) equals approximately 2.1 V. If we subtract this voltage drop from our regulated 5V dc power supply , we can supply approximately 3 V dc to the gearbox motors.Braitenberg Vehicles 133 Figure 9.13 Multidirectional wheel. Figure 9.14 Drawing of U bracket for multidirectional wheel. CdS photoresistor cells As with Walter’s turtletype robot, we use two CdS photoresistor cells . The CdS photoresistors (see Fig. 9.18) used in this robot ha ve a dark resistance of about 100 kV and a light resistance of 10 k V. The CdS photoresistors typically ha ve large variances in resistance between cells. It is useful to use a pair of CdS cells for this robot that matches, as best as one can match them, in resistance. Since the resistance values of the CdS cells can vary so greatly , it’s a good idea to buy a few more than you need and measure the resistances to find a pair whose resistances are c lose. There are a few w ays you can measure the resistance. The simplest method to use a voltohmmeter, set to ohms. Keep the light intensity the same as you measure the resistance . Choose two CdS cells that are closely matched within the group of CdS cells you ha ve.Figure 9.15 Multidirectional wheel and U brac ket attac hed to vehic le base. Figure 9.16 Underside of Braitenberg vehic le showing wheels and gearbox motor drive. 134RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA0 13 12 11 109876321 18 17 CdS Photocell CdS Photocell Sensor 2 Sensor 1 V1 50 kW V1 50 kW D1 1N4002 R2 330 W Q1 2N3904 DC Motor D2 1N4002 R3 330 W Q1 2N3904 DC Motor MCLR' OSC 1 OSC 2 +3 V Vc +3 V Vc VDD VSS 5 416 15U1 14 R1 4.7 kW C1 .1 μF X1 4 MHz +5 V Vc PIC 16F84 C4 10 μF 20 V C5 100 μF 20 V U2 LM2940 D3 1N4002 D4 1N4002 D5 1N4002 +5 V +3 V 6 V+ I 1 2 3 R O C2 .1 μF C3 .1 μF + + Figure 9.17 Schematic of Braitenberg vehicle. 135136 Chapter Nine Figure 9.18 CdS photoresistor cell. The second method involves building a simple PIC 16F84 circuit connected to an LCD displa y. The advantage of this circuit is that you can see the response of the CdS cells under varying light conditions . In addition, you can see the difference in resistance between the CdS cells when they are held under the same illumination. This numeric difference of the CdS cells under exact lighting is used as a fudge factor in the final turtle program. If you just test the CdS cells with just an ohmmeter, you will end up using a larger fudge factor for the robot to operate properly. The schematic for testing the CdS cells is shown in F ig. 9.19. The circuit, built on a PIC Experimenter’ s Board, is shown in F ig. 9.20. The PicBasic Pro testing program follows: ‘CdS cell test‘PicBasic Pro program‘Serial communication 1200 baud true‘Serial information sent out on port b line 0‘Read CdS cell #1 on port b line 1‘Read CdS cell #2 on port b line 7v1 var byte ‘Variable v1 holds CdS #1 information v2 var byte ‘Variable v2 holds CdS #2 information Pause 1000 ‘Allow time for LCD display main: pot portb.1,255,v1 ‘Read resistance of CdS #1 photocell pot portb.7,255,v2 ‘Read resistance of CdS #2 photocell ‘Display information serout portb.0,1,[$fe,$01] ‘Clear the screenBraitenberg Vehicles 137 LCD Display V1 100KW V2 100KW CdS Cell CdS Cell C2 .1μF 50V C3 .1μF 50V SW4 C1 .1μF R1 4.7KW U1 +5V X1 4MHz 416 15 PIC 16F84 5 VSS VDD 17 181 2 3 6 7 8 910 11 12 13 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA014 MCLR' OSC1 OSC2 Serial Line +5V Gnd Figure 9.19 Schematic of test circuit to matc h CdS cells for use in Braitenberg vehic le. pause 25serout portb.0,1,[”CdS 1 = ”]serout portb.0,1,[#v1]serout portb.0,1,[$fe,$c0] ‘Move to line 2pause 5serout portb.0,1,[”CdS 2 = ”]serout portb.0,1,[#v2]pause 100goto mainNotice in Fig. 9.20 that CdS cell 1 is reading 37 and CdS cell 2 is reading 46 under identical lighting . Keep in mind, this is a c losely matched pair of CdS cells.We can use a fudge factor of ±15 points, meaning that as long as the readings between cells vary from each other by ω15 points, the microcontroller will consider them numerically equal. Trimming the sensor array If you are using the Experimenter’ s Board, you can trim and matc h the CdS cells to one another . Doing so allows you to reduce the fudge factor and produces a crisper response from the robot. Typically one CdS cell resistance will be lower than that of the other CdS cell. To the lower resistance CdS cell add a 1k V (or 4.7kV) trimmer potentiometer in series (see F ig. 9.21). Adjust the potentiometer (trim) resistance until the outputs shown on the LCD displa y equal eac h other. Trim the CdS cell under the same lighting conditions in whic h the robot will function. The138 Chapter Nine Figure 9.20 Test circuit built on PIC Experimenter’s Board. LCD Display V1 100KW V2 100KW CdS Cell CdS Cell C2 .1μF 50V C3 .1μF 50V SW4 C1 .1μF R1 4.7KW U1 +5V X1 4MHz 416 15 PIC 16F84 5 VSS VDD 17 181 2 3 6 7 8 910 11 12 13 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA014 MCLR' OSC1 OSC2 Serial Line +5V Gnd 1KW V3 Figure 9.21 Schematic of test circuit with trimmer potentiometer.Braitenberg Vehicles 139 reason for this is that when the light intensity varies from that nominal point to which you’ve trimmed the CdS cell,the responses of the individual CdS cells to changes in light intensity also vary from one another and then are not as closely matched. PIC 16F84 microcontroller The 16F84 microcontroller used in this robot simulates two neurons . Each neuron’s input is connected to a CdS cell. The output of each neuron activates one gearbox motor. In the program I put in a fudge factor, or range, over which the two CdS cells can deviate from one another in resistance readings and still be considered equal. If the robot doesn’t tra vel straight ahead when the two CdS cells are equally illuminated, you can increase the range until it does . PicBasic Compiler program ‘Braitenberg vehicle 1start:pot 1, 255,b0 pot 2, 255,b1 If b0 = b1 then straightif b0 > b1 then leftif b1 > b0 then rightstraight:high 3: high 4goto startleft:b2 = b0 b1 if b2 > 15 then left1 goto straight left1: high 3: low 4 goto startright: b2 = b1 b0 if b2 > 15 then right1 goto straight right1: high 4: lo3 goto start Testing ‘Read CdS cell # 1‘Read CdS cell # 2‘Compare numerical values +/15‘If greater than 15 turn left‘If not go to straight subroutine‘Turn left‘Motor control‘Compare numerical values +/15‘If greater then 15 points‘Turn toward the right‘If not go straight‘Turn right‘Motor control‘Do againThe finished robot is shown in F ig. 9.22. For power I used 4 AA cell batteries. I pointed one CdS cell to the left and the other to the right (see F ig. 9.23). To140 Chapter Nine Figure 9.22 Finished Braitenberg vehicle. Figure 9.23 Closeup of CdS cells mounted in solderless breadboard.Braitenberg Vehicles 141 test the robot’s function, I used a flashlight. Using the flashlight, I was able to steer the mobile platform around by shining the flashlight on the CdS cells . Second Braitenberg Vehicle (Avoidance Behavior) Given the way the robot is currently wired, it is attracted to and steers toward a bright light source . By reversing the wiring going to the gearboxes you can create the opposite behavior. Parts List (1) Microcontroller (16F84) (1) 4.0MHz crystal (2) 22pF caps (1) 0.1mF cap (1) 100mF cap (1) 10mF cap (2) 0.1mF caps (2) 330V, 1/4W resistors (1) 4.7kV, 1/4W resistor (2) CdS photoresistor cells (see text) (2) 100:1 gearbox motors (918D) (2) NPN transistors (2N3904) (5) Diodes (1N4002) (2) 2.25indiameter wheels (1) Multidirectional wheel (1) Voltage regulator (low dropdown voltage 15 V) (LM2940) Miscellaneous items needed inc lude 6in length of 3mm hollow tubing , aluminum 8 in 3 4 in 3 1/8 in thick, 2 solderless breadboards , 3M doublesided tape, battery holder for 4 D batteries , 3in 1/420 threaded rod, and 2 machine screw nuts.This page intentionally left blank.10 Chapter Hexapod Walker Legged walkers are a c lass of robots that imitate the locomotion of animals and insects, using legs. Legged robots have the potential to transverse rough terrains that are impassable by standard wheeled vehic les. It is with this in mind that robotists are developing w alker robots. Imitation of Life Legged walkers may imitate the locomotion style of insects , crabs, and sometimes humans. Biped walkers are still a little rare , requiring balance and a good deal more engineering science than multilegged robots . A bipedal robot walker is discussed in detail in Chap . 13. In this c hapter we will build a sixlegged walker robot. Six Legs—Tripod Gait Using a sixlegged model, we can demonstrate the famous tripod gait used by the majority of legged creatures. In the following drawings a dark circle means the foot is firmly planted on the ground and is supporting the weight of the creature (or robot). A light circle means the foot is not supporting any weight and is movable. Figure 10.1A shows our walker at rest. All six feet are on the ground. From the resting position our w alker decides to move forw ard. To step forw ard, it leaves lifts three of its legs (see F ig. 10.1B, white circ les), leaving its entire weight distributed on the remaining three legs (dark circ les). Notice that the feet supporting the weight (dark circ les) are in the shape of a tripod. A tripod is a very stable weightsupporting position. Our walker is unlikely to fall over. The three feet that are not supporting any weight may be lifted (white circles) and moved without disturbing the stability of the walker. These feet move forward. Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 143144 Chapter Ten Figure 10.1 Sample biological tripod gait. Figure 10.1C illustrates where the three lifted legs move . At this point, the walker’s weight shifts from the stationary feet to the moved feet (see Fig. 10.1D). Notice that the creature’ s weight is still supported by a tripod position of feet. Now the other set of legs moves forw ard and the cyc le repeats. This is called a tripod gait, because a tripod positioning of legs alw ays supports the weight of the w alker. ThreeServomotor Walker Robot The robot we will build is shown in F ig. 10.2. This walker robot is a compromise in design, but allows us to build a sixlegged w alker using just three servomotors. The threeservomotor hexapod walker demonstrates a true tri pod gait. It is not identical to the biological gait we just looked at, but close enough. This legged hexapod uses three inexpensive HS322 (42oz torque) servomotors for motion and one PIC 16F84 microcontroller for brains . The microcontroller stores the program for w alking, controls the three servomotors , and reads the two sensor switc hes in front. The walking program contains subroutines for w alking forward and bac kward, turning right, and turning left. The two switch sensors positioned in the front of the w alker inform the microcontroller of any obstacles in the walker’s path. Based on the feedback from these switch sensors, the walker will turn or reverse to a void obstacles placed in its path. Function The tripod gait I programmed into this robot isn’t the only workable gait. There are other perfectly usable gaits you can develop on your own. Considerthis walking program a working start point.To modify the program, it’s important to understand both the program and robot leg functions. First let’s look at the robot. At the rear of the walker are two servomotors. One is identified as L for the left side, the other as R for the right side . Each servomotor controls both the front and back legs on its side. The back leg is attached directly to the horn of the servomotor. It is capable of swinging the leg forw ard and bac kward. The back leg connects to the front leg through a linkage . The linkage makes the front leg follow the action of the bac k leg as it swings forward and back. The third servomotor controls the two center legs of the w alker. This servomotor rotates the center legs 20 ° to 30° clockwise (CW) or counterc lockwise (CCW), tilting the robot to one side or the other (left or right). With this information we can examine how this legged robot will w alk. Moving Forward We start in the rest position (see F ig. 10.3). As before, each circle represents a foot, and the dark circles show the weightbearing feet. Notice in the rest position, the center legs do not support any weight. These center legs are made to be 1/8 in shorter than the front and bac k legs. In position A the center legs are rotated CW by about 25°from center position. This causes the robot to tilt to the right.The weight distribution is now on the front and back right legs and the center left leg. This is the standard tripod position as described earlier. Since there is no weight on the front and back left legs , they are free to move forw ard as shown in the B position of Fig. 10.3. Hexapod Walker 145 Figure 10.2 Hexapod robot.146 Chapter Ten Figure 10.3 Forward gait for hexapod robot. In the C position the center legs are rotated CCW by about 25° from center position. This causes the robot to tilt to the left.The weight distribution is now on the front and back left legs and the center right leg. Since there is no weight on the front and back right legs, they are free to move forward, as shown in the D position. In position E the center legs are rotated back to their center position. The robot is not in a tilted position so its weight is distributed on the front and back legs. In the F position, the front and bac k legs are moved bac kward simultaneously, causing the robot to move forw ard. The walking cycle can then repeat. Moving Backward We start in the rest position (see Fig. 10.4), as before. In position A the center legs are rotated CW by about 25 ° from center position. The robot tilts to the right. The weight distribution is now on the front and bac k right legs and the center left leg. Since there is no weight on the front and back left legs, they are free to move backward, as shown in the B position of Fig. 10.4.Hexapod Walker 147 Figure 10.4 Backward gait for hexapod robot. In the C position the center legs are rotated CCW by about 25 ° from center position. The robot tilts to the left. Since there is no weight on the front and back right legs, they are free to move backward, as shown in the D position. In position E the center legs are rotated bac k to their center position. The robot is not in a tilted position,so its weight is distributed on the front and back legs. In the F position, the front and bac k legs are moved forw ard simultaneously, causing the robot to move backward. The walking cycle can then repeat. Turning Left The leg motion sequence to turn left is shown in F ig. 10.5. In position A the center legs are rotated CW by about 25° from center position. The robot tilts to the right. The weight distribution is now on the front and bac k right legs and the center left leg. Since there is no weight on the front and back left legs, they are free to move forward, as shown in Fig. 10.4. In the B position, the center legs are rotated CCW by about 25° from center position. The robot tilts to the left. Since there is no weight on the front and back right legs, they are free to move backward, as shown in the C position.148 Chapter Ten Figure 10.5 Turningleft gait for hexapod robot. In position D, the center legs are rotated bac k to their center position. The robot is not in a tilted position, so its weight is distributed on the front and back legs. In position, the left legs moved backward while the right legs moved forward, simultaneously causing the robot to turn left. It typically takes three turning cycles to turn the robot 90°. Turning Right Turning right follows the same sequence as turning left, with the leg positions reversed. ConstructionFor the main body I used a sheet of aluminum 3 in wide 3 9 in long 3 0.032 in thic k. The servomotors are mounted to the front of the body (see Fig. 10.6). The four 11/64indiameter holes a little past halfway down the main body are for mounting the center servomotor . These four holes are offset to the right side. This is necessary to align the servomotor’s horn in the center of the body.1 11/162-9/16 1-3/16 13 9 5/81-1/16 3/4 3/4 5/8 1-1/16 7/8 7/8 2-1/2 3/16 3 PIVOT HOLES FOR LEGS FOUR 11/64 BRACKET HOLES FOR CENTER SERVOMOTOR 1/2 HOLE TO PASS WIRES THROUGH 1/2 SERVOMOTOR HOLE PLACEMENT ALL DIMENSIONS IN INCHES Hexapod Walker 149 Figure 10.6 Diagram of robot base. The bottom two holes are for mounting the pivots for the two bac k legs. Use a punc h to dimple the metal in the center of eac h hole you plan to drill. This will prevent the drill bit from w alking when you drill the hole . If you don’t ha ve a punc h available, use the pointed tip of a nail for a quic k substitute.150 Chapter Ten 2-3/4 3/4 3/4 2 3-3/4 2-3/4 3-1/4 1/4 HOLE 1/4 HOLE 1/16 HOLE (FOR 0-80 SCREWS) BEND 90° BACK LEG (QUAN. 2) FRONT LEG (QUAN. 2) ALL DIMENSIONS IN INCHES Figure 10.7 Diagram of robot legs (front and back). The legs for the robot are made from 1/inwide 3 1/inthick aluminum bar 2 8 stock (see Fig. 10.7). There are four drilled holes needed in the two bac k legs. The three holes that are c lustered together tow ard one end of the leg are for mounting the leg to a servomotor horn. The two 1/in holes allow a 080 screw 16 to pass through.The centered 1/4in hole allows you to remove or attach the servomotor screw that holds the servomotor horn (and leg assembly) to the ser vomotor. Make sure these three holes line up with the holes on the servomotor horn you intend to use. The front legs only need two holes—one for the pivot and the other for the linkage. Also notice that the front legs are 0.25 in shorter than the bac k legs. This compensates for the height of the servomotor mounting horn on the back servomotors where the bac k legs are attac hed. Shortening the front legs makes the robot platform approximately level.90° 90° 13/4 53/4 90° TWIST ALL DIMENSIONS IN INCHES MATERIAL 1/8 3 1/2 3 91/4 ALUMINUM BAR Hexapod Walker 151 Figure 10.8 Diagram of center tilt legs, which are constructed of a single piece of aluminum and are 1/8 in shorter than the front and bac k legs. After the holes are drilled, we need to bend the aluminum bar into shape . Secure the aluminum bar in a vise 2 3/in from the end with the drilled holes . 4 Pressure is applied to bend the aluminum bar at a 90° angle. It’s best to apply pressure at the base of the aluminum bar c lose to the vise. This will bend the leg at a 90° angle, while keeping the lower portion of the leg straight without any bowing of the lower portion. The center legs are made from one piece of aluminum (see F ig. 10.8). The center legs are about 1/8 in shorter than the front and bac k legs when mounted to the robot. So when centered, the legs do not support any weight. These legs are for tilting the robot to the left or right. The legs tilt the robot by rotating the center servomotor approximately ±20°. To produce the center legs , first drill the servomotor horn’ s mounting holes in the center of the 1/in 3 1/in 3 91/in aluminum bar. This should be simi8 2 4 lar to the three c lustered holes you drilled in the bac k legs. Next secure the aluminum bar in a vise . The top of the vise should hold the aluminum bar 3/in from the center of the aluminum bar . Grab the aluminum bar with pliers about 1/2 in above the vise . Keeping a secure grip with the pliers , slowly twist the aluminum bar 90°. Don’t go fast, or you could easily snap the aluminum bar. Repeat the twist on the other side . After the two 90° twists have been made , make the other 90° bend for the legs, as we have done before for the front and bac k legs. Mounting the servomotors The back servomotors are attached to the aluminum body using plastic 632 machine screws and nuts . The reason I used plastic screws is that the plas4152 Chapter Ten tic is a little flexible , allowing the drilled holes to be slightly offcenter from the mounting holes on the servomotor without creating a problem. The legs are attac hed to the servomotor’ s plastic horn. For this I used 080 machine screws and nuts. When you mount the servomotor horn on the servomotor, make sure that eac h leg can swing forw ard and bac kward an equal amount from a perpendicular position. Leg positioning The legs must be positioned accurately, or the walking program will not cause the hexapod robot to walk properly. To aid in this positioning look at Fig. 10.9. The numbers next to the leg positions represent the pulse width output signal for the servomotors. The circuit we will use to control and power the hexapod walker may also be used to adjust the leg positions . A simplified schematic is shown in F ig. 10.10 that is useful for adjusting the legs . This schematic is almost identical to the schematic that will control the robot; the only difference is that the two sensor switches are removed. The leg adjustment program is small;see below for both PicBasic Pro and PicBasic versions. If you decide to buy the PCB board for this robot (Fig. 10.22), you can use the PCB board for this test circuit and program. To align the legs, first disconnect the servomotor horn from the servomotor by unscrewing the center mounting screw from the horn. Once the screw is removed, pull the horn off. Keep the leg attached to the horn.Apply power to the servomotor and connect the control line of the servomotor to RB4. This will center the servomotor’s rotational position. Now reattach the servomotor horn to the servomotor , positioning the leg to be in the center position, as shown in Fig. 10.9. Lock the servomotor horn in place, using the center screw. The leg is now in proper position. By connecting the servomoFigure 10.9 Diagram of leg positions relating to pulse widths.Hexapod Walker 153 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA0 13 12 11 109876321 18 17 Servo Motor Left Servo Motor Right Servo Motor Tilt MCLR’ OSC1 OSC2 VDD VSS 5 416 15U1 14 R1 4.7 KW C1 .1 μF X1 4 MHz +5 V +5 V PIC 16F84 U2 7805 +5 V +5 V 6-9 V+ – I 1 2 3 R O Figure 10.10 Schematic of test circuit. tor control line to pins RB5 and RB6,you can verify the leg’s front and back swing. Adjust the program if necessary to ensure a proper swing . When switching a servomotor from pin to pin,you must power down the circuit first. If you just switch pins without powering down, the microcontroller could latch up and you will get inaccurate positioning . ‘Leg adjustment program (PicBasic Pro)for 16f84 microcontrollerstart: pulsout portb.4, 150 ‘Pin rb4 pulsout portb.5, 120 ‘Pin rb5 pulsout portb.6, 180 ‘Pin rb6 pause 18 goto start end ‘Leg adjustment program (PicBasic)for 16f84 microcontrollerstart:pulsout 4, 150 ‘Pin rb4pulsout 5, 120 ‘Pin rb5pulsout 6, 180 ‘Pin rb6pause 18goto startend154 Chapter Ten PIVOT VIEW A BINDING POST BODY PLASTIC WASHERS SCREW LEG VIEW A ALL DIMENSIONS IN INCHES Figure 10.11 Diagram of robot base with front and bac k leg linkage. View A shows detail of pivot for front legs. Linkage The linkage between the front and bac k legs is made from standard Radio Control (RC) clevis linkage (see Fig. 10.11). In the prototype robot the linkage is 63/4 in center to center. The linkage fits inside the holes in the front and back legs. The back legs must be attached to the body of the robot before you make the linkage. The pivot for the front legs is made from a 3/8in binding post and screw. The leg is attac hed as shown in the c loseup in F ig. 10.11. The plastic washers underneath the body are necessary. They fill up the space between the aluminum body and the bottom of the screw. This keeps the leg close to the aluminum body without sagging. I choose plastic washers for less friction. Do not use so many washers that force is created,binding the leg to the body. The joint should pivot freely. Center (tilt) servomotor To attach the center servomotor to the body requires two Lshaped brac kets (see Fig. 10.12). Drill the holes and bend at a 90° angle.Hexapod Walker 155 Figure 10.12 Closeup of clevis linkage. Attach the two L brackets to the center servomotor, using the plastic screws and nuts (see F ig. 10.13). Next mount the center servomotor assembly under the robot body . Align the four holes in the body with the top holes in the L brackets. Secure with plastic screws and nuts. You must align the center legs on the center servomotor properly, or else the robot will not tilt properly . First remove the horn from the center servomotor . Then attach the center leg to the removed horn, using the 080 screws ands nuts. Apply the center control signal (RB4 from F ig. 10.10) to the center servomotor.With the servomotor centered, reattach the horn/center leg assembly to the servomotor, making sure that the legs are in the center position when securing it in position. Once the center leg is attac hed, you can remove power from the servomotor. Figures 10.14 and 10.15 show the underside and top side of the hexapod robot. Sensors This hexapod has two front switc h sensors for detecting obstac les (see F ig. 10.16). The switch is a miniature snapaction flat lever arm, model number TFCGV3VT185BC manufactured by C&K Components . The levers on the switches are retrofitted with feelers that extend the range of the levers for ward and to the side. The feelers are made with miniature metal tubing or stiff wire (aluminum, steel, or copper).BEND 90° 2 3ALL DIMENSIONS IN INCHES 156 Chapter Ten Figure 10.13 Diagram of L bracket needed for tilt servomotor. Figure 10.14 Tilt servomotor with brackets ready to be attached to robot base. To attach the feelers to the lever , I used a 3/8inlong piece of small rubber tubing. I slid two sections of tubing onto the lever , then slid the stiff wire underneath the tubing (see Fig. 10.17). Attaching the switches to the front of the hexapod required a small fixture to prevent the mounting screws for the switches from getting in the way of the moving front legs . The fixture is made from two pieces of wood. One piece of wood measures 1/in wide 3 1/in thick 3 1 in long. The second piece of wood 2 4 measures 3/in wide 3 1/in thick 3 3 in long. 4 4Hexapod Walker 157 Figure 10.15 Tilt servomotor attached to robot base. Figure 10.16 Snapaction lever switch used for front obstacle sensors. Figure 10.17 Bottom view of switch assembly showing feelers.158 Chapter Ten Figure 10.18 Switch assembly cutaway drawing. Figure 10.18 illustrates the construction of the switc h assembly. The two switches are mounted diagonally on the 3inlong piece of wood using plastic machine screws and nuts. The 1inlong piece of wood is mounted on top of the 3inlong piece of wood. Two holes are drilled through the robotic base and two pieces of wood. The assembly is mounted to the robotic base using two plastic machine screws and nuts. Figures 10.19 and 10.20 show the front and bottom views of the switc h assembly. Electronics Figure 10.21 shows the sc hematic for the servomotors and PIC microcon troller. Notice the 6V battery pac k is powering the microcontroller as well as the servomotors. The battery pack is a 16V unit using four AA batteries. The microcontroller circuit may also be built on a small printedcircuit board that is a vailable from Images SI Inc . (see Fig. 10.22). The robot will function for a short time using a fresh 9V battery , it will deplete quic kly. A secondary battery pack may be laid on top of the aluminum body and connected to the PC board using a power plug.Hexapod Walker 159 Figure 10.19 Front view of switch assembly attached to robot base. Figure 10.20 Bottom view closeup of switch assembly. Figure 10.23 shows the completed w alker ready to run. Microcontroller program The 16F84 microcontroller controls the three servomotors, using just three I/O lines. This leaves 10 available I/O lines and plenty of programming space left over to improve and add to this basic w alker. The program follows: ‘Hexapod walker‘Notes‘Servomotor configuration‘Left leg(s) servomotor connected to rb4‘Right leg(s) servomotor connected to rb5‘Center tilt servomotor connected to rb6‘Pulse width out signals for following servomotors:‘Left leg (150 center) (180 forward) (120 back)‘Right leg (150 center) (120 forward) (180 back)‘Tilt (left 170) (right 130) (center 150)160 Chapter Ten RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA0 13 12 11 109876321 18 17 Servo Motor Left Servo Motor Right Servo Motor Tilt MCLR ’ OSC1 OSC2 VDD VSS 5 416 15U1 14 R1 4.7 KW C1 .1 μF X1 4 MHz +5 V +5 V PIC 16F84 U2 7805 +5 V +5 V +5 V +5 V +5 V 6-9 VDC + – I 1 2 3 R O R2 10 KW R3 10 KW SW2 Left SW1 Right Figure 10.21 Schematic of hexapod circuit. 10 KW 10 KW 10 KW 10 KW 10 KW 10 KW 10 KW 10 KW 10 KW 4.7 KW 470 W + – 1 2 3 4 1 2 3 4 LED Images SI Inc. NYSer Out Gnd U1 16F84 Top9 V Battery Bottom C1 C2 + + – – C1 On Off Reset DC Power Jack PJ-102B Servomotor Controller SW1 Right SW2 Left Tilt Servomotor Right Servomotor Left Servomotor D1 Figure 10.22 Placement of components on stock PC board available from Images SI Inc.Hexapod Walker 161 Figure 10.23 Finished robot. ‘Declare variables ls var byte ‘Left servomotor pulse width rs var byte ‘Right servomotor pulse width cs var byte ‘Center servomotor pulse width ct var byte ‘Count b0 var byte ‘Count b1 var byte ‘Count ‘Define variablesls = 150rs = 150cs = 150pause 250start:‘Read forward sensors‘Front collision?if (porta.1 = 0 && porta.2 = 0) then‘Both left and right sensors are hit, move backwardfor b0= 1 to 3gosub backstepnextfor b0= 1 to 4gosub rturnnextendif‘Collision on right?162 Chapter Ten if (porta.1 = 1 && porta.2 = 0) then‘Right sensor hit, collision on rightfor b0= 1 to 2gosub backstepnextfor b0= 1 to 4gosub lturnnextendif‘Collision on left?if (porta.1 = 0 && porta.2 = 1) then‘Left sensor hit, collision on leftfor b0= 1 to 2gosub backstepnextfor b0 = 1 to 4gosub rturnnextendif‘No collision keep moving forwardif (porta.1 = 1 && porta.2 = 1) then‘Sensors cleargosub frwardendifgoto start‘============================================================‘Function subroutines‘============================================================backstep: ‘Backward stepgosub rtiltgosub rlfgosub ltiltgosub llfgosub centerls = 150: rs = 150gosub do_itreturnfrward: ‘Forward stepgosub rtiltgosub rlbgosub ltiltgosub llbgosub centerls = 150: rs = 150Hexapod Walker 163 gosub do_itreturnlturn: ‘Left turngosub rtiltgosub rlbgosub ltiltgosub llfgosub centerls = 150: rs = 150gosub do_itreturnrturn: ‘Right turngosub rtiltgosub rlfgosub ltiltgosub llbgosub centerls = 150: rs = 150gosub do_itreturn‘=============================================================‘Primary subroutines‘=============================================================do_it: ‘Move robotforward, backward, left or rightfor b1 = 1 to ctpulsout portb.6, cspulsout portb.5, rspulsout portb.4, lspause 18next b1returncenter: ‘Center tilt servomotorct = 15cs = 150gosub do_itreturnrlf: ‘Right leg forwardct = 20rs = 120gosub do_itreturnrlb: ‘Right leg backct = 20164 Chapter Ten rs = 180gosub do_itreturnllf: ‘Left leg forwardct = 20ls = 180gosub do_itreturnllb: ‘Left leg backct = 20ls = 120gosub do_itreturnrtilt: ‘Right side tiltct = 15cs = 130gosub do_itreturnltilt: ‘Left side tiltct = 15cs = 170gosub do_itreturnThis PicBasic program provides for forw ard, backward, turn left, and turn right motions. Two sensors switches on the front of the robot inform the microcontroller when it has encountered an obstac le. When an obstac le is encountered, the robot steps bac k and turns to the left or right, depending on which side the obstacle was encountered. The robot is provided with a righthandedness . If a front collision is detect ed, the robot steps back, then turns to the right and proceeds forw ard. Parts List ServomotorsMicrocontrollers (16F84)PCBAluminum barsAluminum sheetsThreaded rods and nuts (440)Plastic machine screws, nuts, and washersAvailable from Images SI Inc. (see Suppliers at end of book).Chapter 11 Speech Recognition In the near future , speech will be the method for controlling appliances , toys, tools, computers, and robotics. There is a huge commercial market w aiting for this technology to mature. Our speech recognition circuit is a standalone trainable speec h recognition circuit that may be interfaced to control just about anything electrical (see Fig. 11.1). The interface circuit we will build in the second part of this chapter will allow this speec h recognition circuit to control a variety of electrical devices such as appliances, test instruments, VCRs, TVs, and of course robots. The circuit is trained (programmed) to recognize words you w ant it to recognize. The unit can be trained in any language and even nonlanguages suc h as grunts , birdcalls, and whistles. To be able to control and operate an appliance (computer , VCR, TV security system, etc.) or robot by speaking to it makes it easier to work with that device, while increasing the efficiency and effectiveness . At the most basic level, speech commands allow the user to perform parallel tasks (i.e., hands and eyes are busy elsewhere) while continuing to work with the computer , appliance, instrument, or robot. The heart of the circuit is the HM2007 speech recognition integrated circuit (see Fig. 11.2). The chip provides the options of recognizing either 40 words each with a length of 0.96 s or 20 words eac h with a length of 1.92 s . This speech recognition circuit has a jumper setting (jumper WD on main board) that allows the user to c hoose either the 0.96s word length (40word vocabulary) or the 1.92s word length (20word vocabulary). For memory the circuit uses an 8K 3 8 static RAM. There is a backup memory battery for the SRAM on the main board. This battery keeps the trained words safely stored in the SRAM when the main power is turned off . The button battery lasts approximately 2 years.Without the battery backup you would have to retrain the circuit every time the circuit w as switched off. 165 Copyright © 2004 The McGrawHill Companies. Click here for terms of use .166 Chapter Eleven Figure 11.1 Speech recognition circuit assembled. Figure 11.2 HM2007 integrated circuit. The chip has two operational modes: manual mode and CPU mode . The CPU mode is implemented when it is necessary for the c hip to work as a speech recognition coprocessor under a host computer. This is an attractive approach to speech recognition for computers because the job of listening to sound and recognition of command words doesn’t occupy any of the main computer’s CPU time. In one type of programming scenario, when the HM2007 recognizes a command, it can signal an interrupt to the host CPU and then rela y the command it recognized. The HM2007 chip can be cascaded to provide a larger word recognition library . The SR06 circuit we are building operates in the standalone manual mode. As a standalone circuit, the speech recognition circuit doesn’t require a host computer and may be integrated into other devices to add speec h control.Speech Recognition 167 ApplicationsApplications of command and control of appliances and equipment inc lude these: Telephone assistance systems Data entry Speechcontrolled toys Speech and voice recognition security systems Robotics Software Approach Currently most speec h recognition systems a vailable today are softw are programs that run on personal computers . The softw are requires a compatible sound card be installed in the computer . Once activated, this software runs continuously in the background of the computer’s operating system (Windows, OS/2, etc.) and any other application program. While this speec h software is impressive , it is not economically viable for manufacturers to add personal computer systems to control a w ashing machine or VCR. The speec h recognition softw are steals processing power from the operating system and adds to the computer’ s processing tasks . Typically there is a noticeable slowdown in the operation and function of the computer when voice recognition is enabled. Learning to Listen We take our ability to listen for granted. For instance, we are capable of listening to one person speak among several at a party .We subconsciously filter out the extraneous conversations and sound. This filtering ability is beyond the capabilities of today’s speech recognition systems. Speech recognition is not speech understanding. Understanding the meaning of words is a higher intellectual function. The fact that a computer can respond to a vocal command does not mean it understands the command spoken. Voice recognition systems will one da y ha ve the ability to distinguish linguistic nuances and the meaning of words, to “Do what I mean, not what I say!” SpeakerDependent and SpeakerIndependent Recognition Speech recognition is c lassified into two categories , speakerdependent and speakerindependent. Speakerdependent systems are trained by the individual who will be using the system. These systems are capable of achieving a high command count and better than 95 percent accuracy for word recognition. The drawback to this168 Chapter Eleven approach is that the system only responds accurately to the individual who trained the system. This is the most common approac h employed in softw are for personal computers. Speakerindependent systems are trained to respond to a word regardless of who speaks. Therefore the system must respond to a large variety of speec h patterns, inflections, and enunciations of the target word. The command word count is usually lower than that of the speaker dependent system; however, high accuracy can still be maintained within processing limits . Industrial requirements more often require speaker independent voice systems, such as the AT&T system used in the telephone systems . Recognition Style Speech recognition systems ha ve another constraint concerning the style of speech they can recognize. They are three styles of speech: isolated, connected, and continuous. Isolated speech recognition systems can just handle words that are spoken separately. This is the most common speec h recognition system a vailable today. The user must pause between eac h word or command spoken. The speech recognition circuit is set up to identify isolated words of 0.96s length. Connected speec h recognition system is a halfw ay point between isolated word and continuous speec h recognition. It allows users to speak multiple words. The HM2007 can be set up to identify words or phrases 1.92 s in length. This reduces the word recognition vocabulary number to 20. Continuous speech is the natural conversational speec h we are used to in everyday life. It is extremely difficult for a recognizer to sift through the text as the words tend to merge together . For instance, “Hi, how are you doing?” sounds like “Hi, howyadoin.” Continuous speec h recognition systems are on the market and are under continual development. Speech Recognition Circuit The speech recognition circuit is available as a kit from Images SI Inc. You can purchase the main components , HM2007, SRAM, and printedcircuit boards separately if you like and build from scratch. The kit takes a modular approach and uses three separate printedcircuit (PC) boards . The three PC boards are the main circuit board containing the speech recognition circuit, digital display board, and keypad (see Fig. 11.3). The keypad and digital display are removable from the main circuit board. They are needed to communicate with and program the main speec h recognition circuit. After the programming is accomplished, the digital display and keyboard can be removed, and the main circuit embedded into another circuit to add speec h control. Circuit construction The schematic is shown in F ig. 11.4. You can hardwire this circuit to a breadboard if you like . I would recommend purc hasing the three PCB boards thatSpeech Recognition 169 Keypad Display Board Main Circuit Board Figure 11.3 Three modular circuit boards. are available for this project; see Parts List. When you use the PC board, the components are mounted on the top silkscreen side of the board. Begin construction by soldering the IC sockets onto the PC boards. Next mount and solder all the resistors. Now mount and solder the 3.57MHz crystal and red LED. The long lead of the LED is positive. Next solder the capacitors and 7805 voltage regulator. Solder the seven position headers on the keypad to the main circuit board. Next solder the 10 position headers on the display board and main circuit board. Keypad The keypad is made up of 12 normally open (N .O.) pushbutton switc hes (see Fig. 11.5). 1 2 34 5 67 8 9* 0 #Clear TrainTo train To train the circuit, first attach the keypad and digital display to the main circuit board (see Fig. 11.6). Next select your word length. Place a jumper on the two pin WD header on the main circuit board to select a 20word vocabulary , each with a 2s word length. Leave the jumper off to select a 40word vocabulary, each with a 1s word length. Plug in the headset microphone . When power is applied, the HM2007 checks the static RAM, outputs “00” on the digital displa y, and lights the red LED (READ Y). The circuit is in the ready7805 VDC In Vcc +5V Vcc +5V LED XTAL 3.57 MHz R3 22K R2 6.8K .1 μF C2 Microphone Keypad (See Switch Matrix) 1 2 3 4 5 6 7 8 9 CLR 0 TRN HM 2007 46 15 47 44 7 3211 10 9 8 6 5 436 37 38 39 40 41 42 43 31 30 29 28 27 24 23 22 21 20 19 18 17 16 1 12 26 35 34 25 16-Pin Dip Resistor 220 W 16-Pin Dip Resistor 220 W Vcc +5V Vcc +5V C3 100 μF 7448 7448 8K x 8 SRAM 74LS373 23456789 10 27 22 20 23 11 11 14 13 818 17 43 7 12 13 15 15 12 9 6 2 516 19 16 17 18 19 28 26 21 24 25 R1 100K .0047 μF C1 +3V Backup Vcc Header 7 1 2 6 5 4 313 13 8761 214 13 8761 214 12 4 12 4 12 11 10 915 14 13 12 11 10 915 14 7 1 2 6 5 4 3 Figure 11.4 Schematic of speech recognition circuit. 170Speech Recognition 171 Figure 11.5 Keypad wiring. Figure 11.6 Modular components put together for training. mode. In the ready mode the circuit is listening for a verbal command or waiting to be trained. To train the circuit, begin by pressing the word number you want to train on the keypad. In this exercise I am assuming you choose the 20word vocabulary. In this mode the circuit can be trained to recognize up to 20 words . Use any172 Chapter Eleven numbers between 1 and 20. For example, press the number 1 to train word number 1. When you press the number(s) on the keypad,the red LED will turn off. The number pressed on the keypad is shown on the digital displa y. Next press the # key for train. When the # key is pressed, it signals the c hip to listen for a training word, and the red LED turns bac k on. Now speak the word you want the circuit to recognize into the headphone microphone c learly. The LED should blink off momentarily; this is a signal that the word has been accepted. Continue training new words in the circuit, using the procedure outlined above. Press the 2 key, then the # key to train the second word, and so on. The circuit will accept up to either 20 or 40 words, depending on the lengths of the words. You do not have to enter 20 words into memory to use the circuit.If you want, you can use as few word spaces as you require . The procedure for training 40 words is identical, except that you can choose word numbers between 1 and 40. Testing Recognition The circuit is continually listening . Repeat a trained word into the micro phone. The number of the word should be displa yed on the digital display. For instance, if the word directory was trained as word number 5, then saying the word directory into the microphone will cause the number 5 to be displa yed. Error codes The chip provides the following error codes. 55 5 word too long 66 5 word too short 77 5 word no match Clearing the trained word memory To erase all the words in the SRAM memory (training),press 99 on the keypad and then press the * key . The displa y will scroll through the numbers 1 through 20 (or 1 through 40 if in 1s word length mode) quic kly, clearing out the memory. To erase a single word space, press the number of the word you want to clear and then press the * key. Independent Recognition System In addition to speech commands, this circuit allows you to experiment with other facets of speec h recognition tec hnology. For instance , you can experimentSpeech Recognition 173 with speakerindependent systems. This system is inherently speaker dependent, meaning that the voice that trained the circuit also uses it.To experiment with speakerindependent recognition (multiuser), try the following technique. Set the WD jumper on the main circuit board to the 40word vocabulary with a 0.96s word length. Now we will use four word spaces for eac h command word. We will arrange the words so that the command words will be recognized by just decoding the least significant digit (number) on the digital displa y. This is accomplished by allocating the word spaces 01, 11, 21, and 31 to the first target or command word.When the circuit is in recognition mode, we only decode the least significant digit number, in this case X1 (where X is any number from 0 to 3) to recognize the target word. We do this for the remaining word spaces . For instance, the second target word will use word spaces 02, 12, 22, and 32.We continue in this manner until all the words are programmed. If possible, use a different person to speak the word.This will enable the system to recognize different voices , inflections, and enunciations of the target word. The more system resources that are allocated for independent recogni tion, the more robust the circuit will become . There are certain ca veats to be a ware of . First you are trading off word vocabulary number for speaker independence . The effective vocabulary drops from 40 words to 10 words. The speech interface control circuit shown later may be used in this speakerindependent experimental capacity. Voice Security System This HM2007 w asn’t designed for use in a voice security system. But this doesn’t prevent you from experimenting with it for that purpose . You may want to use three or four keywords that must be spoken and recognized in sequence in order to activate a circuit that opens a loc k or allows entry. Speech Interface Control Circuit Okay, you ha ve a functioning speec h recognition circuit, so now what? You need a method of allowing those voice commands to activate other electrical devices or functions. To do this, we need to build a universal speec h interface circuit. When designing this interface, I weighed options that I thought would make this interface useful to as many different users as possible . The first parameter I considered w as how many outputs the interface should ha ve. I decided upon 10 outputs . The second consideration w as the type of output that the interface board should provide . Here was a tough c hoice. I had the option to make the output an active high signal that the user could use to activate or be detected. This output could be used on a TTL logic line or CMOS logic line , or to turn on a transistor switc h or power relay in their circuitry.174 Chapter Eleven The other option I thought of w as to put 10 miniature SPDT rela ys on the interface board. This way the interface board could switc h electric power on and off directly from the board. The advantage of the active high output signal is cost.This board would cost much less than the interface board containing 10 relays. The advantage of the relay board is that the miniature power rela ys have enough current capacity to directly control small dc motors and other electric circuits . I couldn’t decide between the two approac hes, so I ha ve inc luded both designs. You can c hoose whic h interface circuit suits you. The front ends of both circuits are identical and function in the same manner . The outputs are different and are explained separately. Since we are controlling 10 outputs , we only need 11 commands—10 com mands for active on/off switches and 1 command to turn everything off. In general, it is better if the main speech recognition board jumper (WD) is set to the 20 twosecond word length option. The 20 twosecond word mode has a better word recognition accuracy than the 40 onesecond setting. However, the interface board will work with both modes . This makes it possible to experiment with the speakerindependent system described earlier. The speech interface circuit needs to perform a couple of jobs . First it needs to determine when the speec h recognition circuit has detected a spoken word. After a word has been detected, it must distinguish whether the word detected is a recognized command word or an unrecognized word. If the word is a recognized command word, it passes the binary information to the output. If the detected word is not a command word, it must block any change to the output. How the circuit works Before we can get into the nuts and bolts of how the interface circuit functions, we must look at the binary information output by the speec h recognition circuit. The output of the speec h recognition circuit consists of two 4bit binarycoded decimal (BCD) numbers. This binary (BCD) information is shown on the speech circuit’s twodigit digital display. Whenever a word is detected, the circuit uses the digital displa y to output the word number it has recognized, or else it outputs its unrecognized/error code . If the word detected is not recog nized, the circuit will display one of the following error codes: 55 5 word too long 66 5 word too short 77 5 word no match Our interface design incorporates a PIC microcontroller (see F ig. 11.7 or 11.8). A preprogrammed microcontroller’s (16F84) first job is to determine if a word has been spoken. To do this , we use an LM339 comparator . A reference voltage for the comparator is generated using a voltage divider made up ofSpeech Recognition 175 Vcc R4 5.6KW R5 15KW R3 10KW LED Input U3 LM339 U4a 4011 U4b 4011 U4c 4011 U4d 4011 +– 45 3 34 1265 89 2 12 14 10 11 11 14 16 154 9876 5 18 17 10321 12 13Vcc ABCD ABCD GND PIC16F84 RB4 RB5 RA4 RA3 RA2 RA1 RB3 RB2 RB1 RA0 RB0/INT VSS VDD U5 MCLR' OSC1 OSC2 X1 4MHzR11 4.7KW Vcc Vcc 24 11 12 18 19 10 987654321 20 21 22 23 A B C D Q10 Q9 Q8 Q7 Q6 Q5 Q4 Q3 Q2 Q1 Q0 3 2 5 4 7 6 9 10 11 12 14 15 3 2 5 4 7 6 9 10 U6a 4049 U6b 4049 U6c 4049 U6d 4049 U6e 4049 U6f 4049 U7a 4049 U7b 4049 U7c 4049 U7d 4049 12345678910 Output 74154 Figure 11.7 Speech recognition interface (active high outputs) SRI03. resistors R4 and R5. The reference voltage is placed on pin 5 of the comparator. Pin 4 of the comparator is connected to the LED lead on the speech recognition circuit. Whenever a word is detected, the LED blinks off momentarily. The output of the comparator (pin 2) is connected to pin 10 (RB4) of the 16F84 microcontroller. The output of the comparator (pin 2) is usually high (15 V). When a176 Chapter Eleven word is detected, the output (pin 2) drops to ground momentarily . The microcontroller monitors this line to determine when a word has been detected. Once a word has been detected, it is necessary for the interface to read the BCD output from the speec h recognition circuit. By using the highand lowdigit BCD nibbles , it’s possible to distinguish trained target words . To do so , the interface must distinguish the error codes 55, 66, and 77 from trained words numbered 5, 6, and 7. To accomplish this, the interface circuit uses four NAND gates off the 4011 integrated circuit. The NAND gates are connected to the highdigit nibble . If the highdigit BCD nibble has the equivalent word numbers of 5, 6, or 7, the output from the four NAND gates is low. The output from the four NAND gates is connected to pin 11 (RB5) of the 16F84. The 16F84 reads this pin to determine if the highdigit nibble is a 5, 6, or 7 (0 V or ground). If these numbers are not displa yed, the output of the NAND gates is high (15 V). So far our circuit can tell when a word has been detected and if the resulting word is an error code . If the output of the speec h recognition circuit is an error code, nothing else happens; the microcontroller loops bac k to the beginning of the program, waiting for another word detection. On the other hand, if a word is detected and it is not an error code , the microcontroller passes the lowdigit number through to the 74HC154 (4to 16line decoder) IC . The 74HCT154 decoder reads the binary number passed to it and brings the corresponding pin equivalent to that number low. PIC 16F84 microcontroller program The PIC 16F84 used in both interface circuits contains the following PicBasic program: ‘Speech recognition interface programsymbol porta = 5 symbol trisa = 133 symbol portb = 6 symbol trisb = 134 poke trisa, 255 poke trisb, 240 start: peek portb, b0 if bit4 = 0 then trigger ‘Trigger enabled, read speech recognition circuit goto start ‘Repeat trigger: pause 500 ‘Wait .5 second peek portb, b0 ‘Read bcd number if bit5 = 1 then send ‘Output number goto start ‘Repeat send: peek porta, b0 ‘Read port a if bit4 = 1 then eleven ‘Is the number 11 poke portb, b0 ‘Output numberSpeech Recognition 177 goto start ‘Repeat eleven: if bit0 = 0 then ten poke portb, 11 goto start ‘Repeat ten: poke portb, 10 goto start ‘Repeat end Active high output The outputs from the 74HCT154 each pass through a 4049 inverting buffer to supply a 15Vdc active high output signal. SPDT relay output In Fig. 11.8, the front end of the circuit is identical to F ig. 11.7. The changes are seen in the back end of the circuit. The active low output signals from the 74HCT154 each connect to one of the 10 PNP transistors , each of whic h controls a corresponding relay. Each relay has a normally open (N.O.) switch and normally closed (N.C.) switch. The relay switches are rated at 124 V ac at 0.5 A or 24 V dc at 1 A. The relay itself consumes approximately 30 mA of current when it is turned on. Circuit Construction There is nothing critical about the circuit construction. The circuit ma y be wired point to point on a breadboard, if you like. Printedcircuit boards make the construction easier and are a vailable as kits from Images SI Inc . The only component that needs special notice is the 10pin female header. If you are not using the PC boards from the kit, you must follow the sc hematic and wire the 10pin female header exactly; or else the interface will not be receiving the signals it expects, and the unit will fail. Programming the Speech Recognition Circuit: Training, Testing, and Retraining Program the speech recognition circuit per the directions given previously. Choose the words you want to use to control the 10 electrical relays or outputs.To turn off all electrical outputs on the interface, train word number 11 as stop, end, or quit. Before you connect the interface to any circuit, repeat all the trained words into the microphone . The corresponding word number will be dis played on the digital display. You should achieve recognition accuracy of better than 95 percent. If the circuit continually confuses two training words , try retraining one of the words . To retrain a word, press the word number, using the keypad; the word number will be displayed on the digital display.178 Chapter Eleven Vcc R4 5.6KW R5 15KW R3 10KW LED Input U3 LM339 U4a 4011 U4b 4011 U4c 4011 U4d 4011 +– 45 3 34 1265 89 2 12 14 10 11 11 14 16 154 9876 5 18 17 10321 12 13Vcc ABCD ABCD GND PIC16F84 RB4 RB5 RA4 RA3 RA2 RA1 RB3 RB2 RB1 RA0 RB0/INT VSS VDD U5 MCLR' OSC1 OSC2 X1 4MHzR11 4.7KW Vcc Vcc 24 11 12 18 19 10 987654321 20 21 22 23 A B C D Q10 Q9 Q8 Q7 Q6 Q5 Q4 Q3 Q2 Q1 Q0 74154 RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW RelayVcc 2N3906 1N4002 100KW Figure 11.8 Speech recognition interface (relay switch outputs) SRI02. Press the T (training) key, and say the word into the microphone. If the circuit still confuses the two words , you may have to c hange one of the suggested words. Once you are satisfied with the accuracy , remove the digital displa y board and the keypad. Next connect the speech interface board to the 10pin header used for the digital display, and you’re ready to go.Speech Recognition 179Figure 11.9 Finished speech recognition board SRI02. Figure 11.10 Finished speech recognition board SRI03.180 Chapter Eleven SRI02 and SRI03 Interfaces The SRI02 and SRI03 built from kits a vailable from Images SI Inc . are shown in F igs. 11.9 and 11.10, respectively. Once the speec h recognition circuit is programmed, the speech recognition interfaces ma y be plugged into the displa y board output on the main speec h recognition board and used. Figure 11.11 shows the SRI02 connected to the speec h recognition board, and Fig. 11.12 shows the SRI03 connected to the speech recognition board. Robot Control The speech recognition circuit uses a headphone microphone. For mobile operation one needs to add a wireless microphone. There are a number of methods of implementing wireless control. The simplest method is to add a suitable microphone to the main circuit board and acoustically couple it to the output of a radio receiver or w alkietalkie. You would use the matc hing w alkietalkie to give voice commands . When using this method, you should train the circuit by using your w alkietalkies and acoustic coupling. Figure 11.11 SRI02 connected to speech recognition circuit.Speech Recognition 181 Figure 11.12 SRI03 connected to speech recognition circuit. Parts List Speech recognition kit (SR06) (1) Speech recognition IC (HM2007) (1) 8K static RAM (6264) (1) Octal latch (74LS373) (1) Display chip (74LS48) (1) 3.57MHz crystal (12) PCmounted N.O. switches (2) Sevensegmented displays (MAN74) (1) Headset microphone (1) 9V battery clip (1) Coin battery holder (2032) (1) PCmounted microphone jack (1) 22kV, 1/W resistor 4 (1) 6.8kV, 1/W resistor 4182 Chapter Eleven (1) 330V, 1/W resistor 4 (8) 220V, 1/W resistor 4 (1) 100kV, 1/W resistor 4 (1) 0.1mf capacitor (1) 100mf capacitor (1) 0.0047mf capacitor (2) 10to 22pF capacitor (1) Voltage regulator (7805) (1) LED (2) 1N914 diode Miscellaneous items needed include PC boards, IC sockets, headers (male and female), twoand threepin connectors, jumpers. Speech interface kit (SRI02) (1) 5.6kV, 1/W resistor 4 (1) 15kV, 1/W resistor 4 (1) 10kV, 1/W resistor 4 (10) 100kV, 1/W resistor 4 (10) Diodes (1N4002) (1) Comparator (LM339) (1) 4011 CMOS NAND (1) 74154 IC (1) PIC 16F84 microcontroller* (10) Omron G5V1 relays Miscellaneous items needed inc lude PC board, 10pin female header, 9V battery clips, and a 7805 regulator. Speech interface kit (SRI03) (1) 5.6kV, 1/W resistor 4 (1) 15kV, 1/W resistor 4 (1) 10kV, 1/W resistor 4 (10) 100kV, 1/W resistor 4 *Preprogrammed 16F84 available separately for $10.00 from Images SI Inc .Speech Recognition 183 (10) Diodes (1N4002) (1) Comparator (LM339) (1) 4011 CMOS NAND (1) 74154 IC (1) PIC 16F84 microcontroller* (2) Inverting buffers (4049) Miscellaneous items needed inc lude PC board, 10pin female header, 9V battery clips, and a 7805 regulator. Speech recognition and interface kits (all components inc luding preprogrammed 16F84 and PCB) available from Images SI Inc. (see Suppliers at end of book): Speech recognition kit (SR06) $79.95 Speech interface kit (SRI03) $89.95 Speech interface kit (relay) (SRI02) $159.95 *Preprogrammed 16F84 available separately for $10.00 from Images SI Inc .This page intentionally left blank.Chapter 12 Robotic Arm Servomotor Building Blocks for Robotics The servomotor brac kets discussed in this c hapter will allow you to create various servomotor robots and projects. Servomotors are ideal for powering robots . They are readily a vailable in many sizes , are inexpensive , provide powerful torque for their size and weight, and are positional. The output shafts on most hobby servomotors are guaranteed positional between 0° and 90°. Most servomotors’ output shaft range extends past 90°, coming close to 180°. The servomotor bracket components are shown in Fig. 12.1. Each of the aluminum U brac kets that make up the assembly has multiple holes for connecting a standard HiTec servomotor horn as well as bottom and top holes for connecting U brackets and assemblies to one another. The servomotor horns used on these servomotor brackets are included with all the compatible HiT ec servomotors, such as HS322, HS425, HS475, and HS35645. These brackets may also be used with similar size Futaba servomotors, but you may have to purchase the horns separately. Each servomotor bracket assembly consists of the following components:two aluminum U brackets, labeled A and B, one binding head post screw, four 632 plastic machine screws with nuts, and four sheet metal screws for mounting a servomotor horn. When assembled with a compatible servomotor (see F ig. 12.2), the bracket becomes a modular component that may be attached to other brackets and components . The bracket allows the top and bottom components to swivel along the axis of the servomotor’ s shaft (see Fig. 12.3). By connecting multiple servomotors using the brac kets, you can create a variety of robotic designs. In this chapter we will use the brackets to create a Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 185186 Chapter Twelve B A Figure 12.1 Servomotor bracket kit. Tilts Front View Side View Figure 12.2 Front and side views of servomotor brac ket. fiveservomotor robotic arm. In Chap. 13 we use these same brac kets to cre ate a bipedal walker robot. The bottom and top have multiple holes for attaching other brackets or servomotor horns (see Fig. 12.4). Basic Servomotor Bracket Assembly To assemble a servomotor brac ket, begin by placing the binding post through the back hole on part a (see Fig. 12.5). Next place servomotor into the A bracket, as shown in F ig. 12.6. Attach the servomotor using 632 3 3/8inlong machine screws and nuts (see Fig. 12.7). Notice the servomotor’s horn has beenRobotic Arm 187 Tilt Left Center Tilt Right Figure 12.3 Servomotor bracket travel. Bracket Holes Horn-Mounting Holes Bracket-to-Bracket Holes Figure 12.4 Diagram of top and bottom mounting holes in the A and B brackets. Figure 12.5 A brac ket with binding screw.removed from the servomotor. To secure the screws at the bottom two positions of the servomotor , place the screw through the hole from the inside of the bracket. It helps if you ha ve a small screwdriver to hold the screw in place . Then the plastic nuts are c hased down on the screws from the outside of the bracket (see Fig. 12.7). The servomotor horn (see F ig. 12.8), is attached to the side holes on the B bracket (see Fig. 12.9). 188 Chapter Twelve Figure 12.6 Side view of placing servomotor in A bracket. Figure 12.7 A brac ket with servomotor attac hed with plastic screws and nuts.Robotic Arm 189 Figure 12.8 HiTec servomotor horn. Back Front Figure 12.9 B bracket with servomotor horn attached. To place the servomotor secure in bracket A into its mating part bracket B, slip the end of the bindingheld post through the hole in the mating part (see Fig. 12.10). Next slip the servomotor’ s spindle into the horn (see F ig. 12.11). Finished assembly is shown in F ig. 12.12. Assembling MultipleServomotor Assemblies When you are using multipleservomotor assemblies, it is essential to preplan how the servomotors will be connected. When two or more servomotors assemblies are connected, the connecting brackets of the joints should be preassembled (see Fig. 12.13). The brackets may be orientated to one another in a number of ways, depending upon your design. The top and bottom brac kets of eac h assembly are connected to one another by four 632 3 3/8inlong plastic machine screws and eight plastic hex nuts. The screws are inserted though the top brac ket holes. Hex nuts190 Figure 12.10 Side view showing horn assembly connected Chapter Twelve Bringing top brac ket onto lower brac ket to assemble. Figure 12.11 to servomotor.Robotic Arm 191Figure 12.12 Standalone servomotor bracket assembly. Figure 12.13 Two different bracket assemblies.192 Chapter Twelve Figure 12.14 Closeup top view of two assembled brac kets. Figure 12.15 Closeup side view of two assembled brac kets. are chased down, securing the machine screws to the top bracket. The second bracket is then attached to the screws, and hex nuts are chased down, securing the bottom bracket. Figures 12.14 and 12.15 are closeup pictures of the top and side views of the plastic screws connecting two brac kets. Building a FiveServomotor Robotic Arm Aside from the servomotor brac kets we ha ve already outlined, we need one other specialized component—a robotic arm gripper (see F ig. 12.16). This gripper requires two servomotors , one for wrist movement and the other to open and c lose the gripper fingers . The gripper fingers can accommodate objects up to about 1.0 in (25 mm).Robotic Arm 193 Figure 12.16 Robotic arm gripper. The robotic arm uses five servomotors: four HiTec HS322 HD servomotors and one HS475 HB servomotor . The HS475 servomotor has 50 percent more torque than the HS322 and is used in the second position up from the bottom (or base) servomotor on the robotic arm. This particular servomotor requires the greatest torque in order to lift the arm and any object the arm is holding. Figure 12.17 shows how the servomotors are attac hed to the gripper . Assemble one part A and B brac ket, as shown in F ig. 12.18. Attach a servomotor to the A portion of the brac ket; this will be the wrist servomotor . The wrist servomotor motor is attac hed to the gripper first. Remove the servomotor horn from the servomotor , if you ha ven’t done so already , and put the horn screw to the side; we will need it. Center the wrist servomotor, using the centering servomotor circuit described later in this chapter or at the end of Chap . 6. With power applied to the servomotor from the centering circuit, place the servomotor into the wrist position. Replace the horn screw removed earlier , and tighten the servomotor horn screw . Remove power from the servomotor. Next position the gripper fingers in midposition. Center the finger servomotor, using the centering circuit as before. Position the finger servomotor in the finger position. Tighten the horn servomotor screw , then bac k off the screw to unbind the fingers . When you are finished, the subassembly should look like Fig. 12.19. To finish up the arm, assemble an A and B component, as shown in F ig. 12.20. Next we require two more A bracket components. One A bracket component has a servomotor horn attac hed to its bottom holes , and the other A bracket component has a servomotor attached and is laid on its back as a base (see Fig. 12.21). The two brackets are assembled as shown in Fig. 12.22. When you assemble the base , center the bottom servomotor before attac hing the upper A bracket. This forms the base of the robotic arm. To secure the base to a platform, four holes are drilled in the bottom brac ket (see Fig. 12.23). Only two drill locations are shown on the bottom. Drill two similar holes at the top. To prevent the A brac ket from bending with the weight of the robotic arm when it is assembled, place a spacer made of wood, plastic, or metal as shown in Fig. 12.23. The base assembly is secured to a square piece of wood or met194 Figure 12.18 Wrist Servo Finger Servo Figure 12.17 Diagram showing how servomotor assembles to Chapter Twelve Assembled brac kets for gripper. gripper.Robotic Arm 195 Figure 12.19 Robotic arm gripper assembly. Figure 12.20 Assembled middle bracket for robotic arm.196 Chapter Twelve Figure 12.21 Bottom brackets for robotic arm. Figure 12.22 Assembled bottom brackets for robotic arm.Spacer Drill Holes Robotic Arm 197 Figure 12.23 Closeup base bracket. al to provide a good base that doesn’t topple when the robotic arm moves and lifts objects. The two middle servomotors are assembled onto the base , and the servogripper is attached to the top, completing the robotic arm (see Figs. 12.24 and 12.25). ServomotorsServomotors are relatively easy to control using PIC microcontrollers . If you remember, servomotors were introduced in Chap . 6. In Chap . 6 we just described the basic function of a servomotor; now we will review in a little greater detail. Servomotors are geared dc motors with a positional feedbac k control that allows the shaft (rotor) to be rotated and positioned accurately . When a control signal is being fed to the servomotor, the servomotor’s shaft rotates to the position specified by the control signal. The positioning control is a dynamic feedback loop, meaning that if you forcibly rotate the servomotor’s shaft away from its control signal command position, the servomotor circuitry will read this as a position error and will increase its torque in an attempt to rotate the shaft back to its command position. Hobby servomotor specifications usually state that the shaft can be posi tioned through a minimum range of 90 ° (±45°). In reality this range can be extended c loser to 180° (±90°) by adjusting the position control signal described in a moment. There are three wire leads to a hobby servomotor . Two leads are for power 15 V (red wire) and ground (black wire). The third lead (yellow or white wire) feeds a position control signal to the motor.198 Figure 12.24 robotic arm (left view). Figure 12.25 robotic arm (right view). Chapter Twelve Fiveservomotor FiveservomotorRobotic Arm 199 Pulse Width 1-2 ms (Approx. Frequency 55 Hz) Period 18 ms 1-ms Pulse Train Servomotor Position Left 1.5-ms Pulse Train Servomotor Position Midrange 2-ms Pulse Train Servomotor Position Right Figure 12.26 Servomotor control signal diagram. The position control signal is a single variablewidth pulse. The pulse width typically varies between 1 and 2 ms. The width of the pulse controls the position of the servomotor shaft. Figure 12.26 illustrates the relationship of pulse width to servomotor position. A 1ms pulse rotates the shaft to the extreme counterclockwise (CCW) position (245°). A 1.5ms pulse places the shaft in a neutral midpoint position (0°). A 2ms pulse rotates the shaft to the extreme CW position (145°). The pulse width signal is sent to the servomotor approximately 55 times per second (55 Hz). By extending our pulse width past the typical parameters , a 1to 2ms pulse width, we can extend the rotational position of the servomotor’ s shaft. In many cases close to 180° positioning control is possible. However, care must be exercised not to provide a control signal to the servomotor that will attempt to rotate the shaft too far, where the shaft will push against its internal stop. As mentioned previously, the position feedbac k control is dynamic , and the servomotor will increase its torque (and increase its current consumption) to rotate the shaft into position, placing as much force as possible against its internal stop . This will create unnecessary strain on the internal gears and motor, decreasing its working life considerably. Servomotor controllers Our servomotor controllers use the PicBasic and PicBasic Pro pulsout command. The command format is as follows:200 Chapter Twelve pulsout pin, periodThe pulsout command generates a pulse on the pin specified for the period of time specified.The time is in 10ms (microsecond) increments. So to send a 1.5ms pulse out on port B pin 0, you could use one of the following command(s). For the PicBasic compiler: pulsout 0, 150For the PicBasic Pro compiler: pulsout portb.0, 150This pulsout command will put the servomotor shaft into its center position. The only things missing are a dela y and loopbac k lines to send the pulsout signal to the servomotor 55 times per second. So a complete center servomotor program is as follows: PicBasic program PicBasic Pro program start: start: pulsout 0, 150 pulsout portb.0, 150 pause 18 pause 18 goto start goto start The schematic for a basic servomotor circuit is shown in F ig. 12.27. If you prototype servomotor circuits on a solderless breadboard, a servomotor connector (see F ig. 12.28) makes connecting a servomotor to the breadboard easy. Although this centering servomotor circuit ma y appear to be useless , it is not. In most cases when building a servomotor device or robot, you want to center the servomotor to a known (center) position before attaching any hardware. This centering tec hnique is used before attac hing the wheel assembly to the steering servomotor when you are constructing Walter’s turtle (see Chaps. 8 and 10 among others). Simple servomotor controller This second servomotor circuit (see Fig. 12.29), allows us to control the servomotor by using a singlepole doublethrow (SPDT) switc h. This particular SPDT switch has a center off position that is critical to proper operation of this circuit. Pushing the switc h up will rotate the servomotor in a c lockwise rotation. In the center position the servomotor stops and holds its position. Pushing the switc h in the down position will rotate the servomotor in the counterclockwise direction. The following two programs for the simple servomotor controller are the basis for the programming for the four and fiveservomotor controllers . In general, when you are programming the PIC microcontrollers, make sure the watchdog timer is disabled.Robotic Arm 201+5V +5V Servo Motor 145 13 12 11 18 17 109876321 16 15 4 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA0 VSS VDD MCLR' OSC1 OSC2 PIC 16F84X1 4MHz U1 R1 4.7KW C1 .1μF Figure 12.27 Centering the servomotor controller circuit. Figure 12.28 Servomotor connector useful for prototyping on solderless breadboards .202 Chapter Twelve +5V +5V +5V Servo MotorSPDT Center Off Switch SW1 SW2 R2 10kW R3 10kW PIC 16F84 13 14 16 15 12 11 10987 63 5 21 18 17 RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/ITOCKI RA3 RA2 RA1 RA0 VSS VDDMCLR' OSC1 OSC2 4 U1 R1 4.7kW C1 .1μF X1 4MHz Figure 12.29 Primary servomotor controller circuit. ‘PicBasic Pro program‘Manual control of servomotor using SPDT switch‘Use b1 to hold pulse width variable for servo 1‘Declare variablesb1 var byte‘Initialize variablesb1 = 150 ‘Start servo 1 at center positionstart:‘Output servomotor positionpulsout portb.0, b1 ‘Send current servo 1 position out‘Check for switch closuresif porta.0 = 0 then left1 ‘Is sw1 left active?if porta.1 = 0 then right1 ‘Is sw1 right active?‘Routine to adjust pause value (nom 18) to generate approx 50 Hz updatepause 18goto startRobotic Arm 203 ‘Routines for servomotor 1 left1: b1 = b1 + 1 ‘Increase the pulse width if b1 > 254 then max1 ‘Maximum 2.54 milliseconds goto start right1: b1 = b1 1 ‘Decrease the pulse width if b1 < 75 then min1 ‘Minimum .75 millisecond goto start max1:b1 = 254 ‘Cap max b1 at 2.54 milliseconds goto start min1:b1 = 75 ‘Cap min b1 at .75 millisecond goto start ‘PicBasic program ‘Manual control of servomotor using SPDT switch‘Use b1 to hold pulse width variable for servo 1‘Declare variables ‘Initialize variables symbol porta = 6 b1 = 150 ‘Start servo 1 at center position start: ‘Output servomotor position pulsout 0, b1 ‘Send current servo 1 position out ‘Check for switch closures peek porta, b0 if bit0 = 0 then left1 ‘Is sw1 left active? if bit1 = 0 then right1 ‘Is sw1 right active? ‘Routine to adjust pause value (nom 18) to generate approx 55 Hz updatepause 18 goto start ‘Routines for servomotor 1 left1: b1 = b1 + 1 ‘Increase the pulse width if b1 > 254 then max1 ‘Maximum 2.54 milliseconds goto start right1:204 Chapter Twelve b1 = b1 1 ‘Decrease the pulse width if b1 < 75 then min1 ‘Minimum .75 millisecond goto start max1:b1 = 254 ‘Cap max b1 at 2.54 milliseconds goto start min1:b1 = 75 ‘Cap min b1 at .75 millisecond goto start Fourand FiveServomotor Controllers The previous schematic is the basic building bloc k used in the fourand fiveservomotor controller. Figure 12.30 shows the fourservomotor controller. This may be purchased as a kit from Images SI Inc ., or you can hardwire the circuit and program the chip yourself. ‘PicBasic Pro program‘Manual control of four servomotors using 4 SPDT switches‘Microcontroller PIC 16f84‘Declare variablesb0 var word ‘Variable for pause routine.b1 var byte ‘Use b1 to hold pulse width variable for servo 1b2 var byte ‘Use B2 to hold pulse width variable for servo 2RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/ITOCKI RA3 RA2 RA1 RA0 13 12 11 109876 321 18 17 VSS 5 14 VDD MCLR' OSC1 OSC2 4 16 15 U1 R1 4.7KW C1 .1μF X1 4MHz +5V +5V +5V +5V +5VServo Motor 1 Servo Motor 2 Servo Motor 3 Servo Motor 4 +5V +5V +5V +5V R9 10KW R8 10KW R7 10KW R6 10KW R4 10KW R5 10KW R2 10KW R3 10KW SW4 SW3 SW2 SW1 PIC 16F84 Figure 12.30 Schematic of fourservomotor controller.Robotic Arm 205 b3 var byte ‘Use b3 to hold pulse width variable for servo 3 b4 var byte ‘Use b4 to hold pulse width variable for servo 4 b5 var byte ‘Variable for pause routine ‘Initialize servomotor variablesb1 = 150 ‘Start up position servo 1b2 = 150 ‘Start up position servo 2b3 = 150 ‘Start up position servo 3b4 = 150 ‘Start up position servo 4start:‘Output servomotor positionpulsout portb.7, b1 ‘Send current servo 1 position outpulsout portb.6, b2 ‘Send current servo 2 position outpulsout portb.5, b3 ‘Send current servo 3 position outpulsout portb.4, b4 ‘Send current servo 4 position out‘Check for switch closuresif porta.0 = 0 then left1 ‘Is sw1 left active?if porta.1 = 0 then right1 ‘Is sw1 right active?if porta.2 = 0 then left2 ‘Is sw2 left active?if porta.3 = 0 then right2 ‘Is sw2 right active?if portb.0 = 0 then left3 ‘Is sw3 left active?if portb.1 = 0 then right3 ‘Is sw3 right active?if portb.2 = 0 then left4 ‘Is sw4 left active?if portb.3 = 0 then right4 ‘Is sw4 right active?‘Routine to adjust pause value (nom 18) to generate approx 50 Hz updateb0 = b1 + b2 + b3 + b4b5 = b0/100b0 = 15 b5pause b0goto start‘Routines for servomotor 1left1:b1 = b1 + 1 ‘Increase the pulse widthif b1 > 254 then max1 ‘Maximum 2.54 millisecondsgoto startright1:b1 = b1 1 ‘Decrease the pulse widthif b1 < 75 then min1 ‘Minimum .75 millisecondgoto startmax1:206 Chapter Twelve b1 = 254 goto startmin1:b1 = 75 goto start‘Routines for servomotor 2left2:b2 = b2 + 1 if b2 > 254 then max2 goto startright2:b2 = b2 1 if b2 < 75 then min2 goto startmax2:b2 = 254 goto startmin2:b2 = 75 goto start‘Routines for servomotor 3left3:b3 = b3 + 1 if b3 > 254 then max3 goto startright3:b3 = b3 1 if b3 < 75 then min3 goto startmax3:b3 = 254 goto startmin3:b3 = 75 goto start‘Routines for servomotor 4left4:b4 = b4 + 1 if b4 > 254 then max4 goto startright4:b4 = b4 1 if b4 < 75 then min4 goto startmax4:b4 = 254 goto startmin4:‘Cap max b1 at 2.54 milliseconds‘Cap min b1 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b2 at 2.54 milliseconds‘Cap min b2 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b3 at 2.54 milliseconds ‘Cap min b3 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b4 at 2.54 millisecondsRobotic Arm 207 b4 = 75 ‘Cap min b4 at .75 millisecondgoto startendFigure 12.31 is a photograph of the completed fourservomotor kit. The circuit board for this kit w as used as the main circuit board for the turtle robot in Chap. 8. Figure 12.32 is a schematic for the fiveservomotor controller. This circuit is suitable for controlling our fiveservomotor robotic arm. When you program the 16F873 with the fiveservomotor controller program, make sure the w atchdog timer is disabled and the brownout reset is Figure 12.31 Assembled fourservomotor controller kit. RB7 RB6 RB5 RB4 RB3 RC3 RC6 RC2 RC1 RC0 RA4 RA5 RA3 RA2 RA1 RA0 28 27 26 25 24 14 13 12 11 765432 VSS 8 1917 20 VDD MCLR' OSC1 OSC2 1 910 U1 R1 4.7KW C1 .1μF X1 4MHz +5V +5V +5V +5V +5VServo Motor 1 Servo Motor 2 Servo Motor 3 +5VServo Motor 4 Servo Motor 5 +5V +5V +5V +5V R11 10KW R10 10KW R9 10KW R8 10KW R4 10KW R5 10KW R2 10KW R3 10KW SW1 SW2 SW4 +5V R7 10KW R6 10KW SW3 SW5 PIC 16F873 Push Button VccR12 10KW Figure 12.32 Schematic of fiveservomotor controller.208 Chapter Twelve also disabled. If the brownout reset is not disabled, the circuit may automatically reset whenever a servomotor draws enough current to make the supply voltage dip momentarily. This is not what you w ant to happen in the middle of a robotic arm operation, so make sure that configuration bit is disabled. These configuration bits are easy to set when you use the EPIC Programmer. Simply go to the Configuration pulldown menu and disable these options . ‘PicBasic Pro program for fiveservomotor controller‘Manual control of five servomotors using 5 SPDT switches‘Microcontroller PIC 16f873adcon1 = 7 ‘Set port a to digital I/O ‘Declare variables b0 var byte ‘Use b0 as hold pulse width variable for servo 1 b1 var byte ‘Use b1 to hold pulse width variable for servo 2 b2 var byte ‘Use b2 to hold pulse width variable for servo 3 b3 var byte ‘Use b3 to hold pulse width variable for servo 4 b4 var byte ‘Use b4 to hold pulse width variable for servo 5 b6 var byte ‘Variable for pause routine b7 var word ‘Variable for pause routine ‘Initialize servomotor variablesb0 = 150 ‘Start up position servo 1b1 = 150 ‘Start up position servo 2b2 = 150 ‘Start up position servo 3b3 = 150 ‘Start up position servo 4b4 = 150 ‘Start up position servo 5start:‘Output servomotor positionportb = 0 ‘Prevents potential signal inversion on resetpulsout portb.7, b0 ‘Send current servo 1 position outpulsout portb.6, b1 ‘Send current servo 2 position outpulsout portb.5, b2 ‘Send current servo 3 position outpulsout portb.4, b3 ‘Send current servo 4 position outpulsout portb.3, b4 ‘Send current servo 5 position out‘Routine to adjust pause value (nom 18) to generate approx 50 Hz updateb7 = b0 + b1 + b2 + b3 + b4b6 = b7/100b7 = 15 b6pause b7‘Check for switch closuresif portc.3 = 0 then left1 if portc.2 = 0 then right1 if portc.1 = 0 then left2 ‘Is sw1 left active?‘Is sw1 right active?‘Is sw2 left active?Robotic Arm 209 if portc.0 = 0 then right2 ‘Is sw2 right active? if porta.5 = 0 then left3 ‘Is sw3 left active? if porta.4 = 0 then right3 ‘Is sw3 right active? if porta.3 = 0 then left4 ‘Is sw4 left active? if porta.2 = 0 then right4 ‘Is sw4 right active? if porta.1 = 0 then left5 ‘Is sw5 left active? if porta.0 = 0 then right5 ‘Is sw5 right active? goto start‘Routines for servomotor 1left1:b0 = b0 + 1 if b0 > 254 then max0 goto startright1:b0 = b0 1 if b0 < 75 then min0 goto startmax0:b0 = 254 goto startmin0:b0 = 75 goto start‘Routines for servomotor 2left2:b1 = b1 + 1 if b1 > 254 then max1 goto startright2:b1 = b1 1 if b1 < 75 then min1 goto startmax1:b1 = 254 goto startmin1:b1 = 75 goto start‘Routines for servomotor 3left3:b2 = b2 + 1 if b2 > 254 then max2 goto startright3:b2 = b2 1 if b2 < 75 then min2 ‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b1 at 2.54 milliseconds‘Cap min b1 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b1 at 2.54 milliseconds‘Cap min b1 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond210 Chapter Twelve goto start max2: b2 = 254 ‘Cap max b2 at 2.54 milliseconds goto start min2: b2 = 75 ‘Cap min b2 at .75 millisecond goto start ‘Routines for servomotor 4 left4: b3 = b3 + 1 ‘Increase the pulse width if b3 > 254 then max3 ‘Maximum 2.54 milliseconds goto start right4: b3 = b3 1 ‘Decrease the pulse width if b3 < 75 then min3 ‘Minimum .75 millisecond goto start max3: b3 = 254 ‘Cap max b3 at 2.54 milliseconds goto start min3: b3 = 75 ‘Cap min b3 at .75 millisecond goto start ‘Routines for servomotor 5 left5: b4 = b4 + 1 ‘Increase the pulse width if b4 > 254 then max4 ‘Maximum 2.54 milliseconds goto start right5: b4 = b4 1 ‘Decrease the pulse width if b4 < 75 then min4 ‘Minimum .75 millisecond goto start max4: b4 = 254 ‘Cap max b4 at 2.54 milliseconds goto start min4: b4 = 75 ‘Cap min b4 at .75 millisecond goto start end Figure 12.33 is a photograph of the fiveservomotor controller . The robotic arm servomotors can plug right onto the three position headers on the main board. However, to separate the control board from the robotic arm, I used five 24in servomotor extensions . Once wired, each SPDT switch controls one robotic arm servomotor (see F ig. 12.34). When using the robotic arm, I noticed the arm move too quic kly for me to perform fine movements. So to slow it down, I added a delay routine. This following program is identical to the above program, with the exception of the delay routine(s).Robotic Arm 211Figure 12.33 Assembled fiveservomotor controller kit. Figure 12.34 Finished robotic arm and fiveservomotor controller.212 Chapter Twelve ‘Slowspeed‘Manual control of five servomotors using 5 SPDT switches‘Microcontroller PIC 16F873adcon1 = 7 ‘Set port a to digital I/O‘Declare variablesb0 var byte ‘Use b0 as hold pulse width variable for servo 1b1 var byte ‘Use b1 to hold pulse width variable for servo 2b2 var byte ‘Use b2 to hold pulse width variable for servo 3b3 var byte ‘Use b3 to hold pulse width variable for servo 4b4 var byte ‘Use b4 to hold pulse width variable for servo 5b6 var byte ‘Variable for pause routineb7 var word ‘Variable for pause routines1 var byte ‘Unassigned delay variables2 var byte ‘Assigned delay variable‘Initialize servomotor variablesb0 = 150 ‘Start up position servo 1b1 = 150 ‘Start up position servo 2b2 = 150 ‘Start up position servo 3b3 = 150 ‘Start up position servo 4b4 = 150 ‘Start up position servo 5s2 = 4 ‘Delay variablestart:‘Output servomotor positionportb = 0 ‘Prevents potential signal inversion on resetpulsout portb.7, b0 ‘Send current servo 1 position outpulsout portb.6, b1 ‘Send current servo 2 position outpulsout portb.5, b2 ‘Send current servo 3 position outpulsout portb.4, b3 ‘Send current servo 4 position outpulsout portb.3, b4 ‘Send current servo 5 position out‘Routine to adjust pause value (nom 18) to generate approx 50 Hz updateb7 = b0 + b1 + b2 + b3 + b4b6 = b7/100b7 = 15 b6pause b7‘Check for switch closuresif portc.3 = 0 then left5 ‘Is sw1 left active?if portc.2 = 0 then right5 ‘Is sw1 right active?if portc.1 = 0 then left4 ‘Is sw2 left active?Robotic Arm 213 if portc.0 = 0 then right4 ‘Is sw2 right active? if porta.5 = 0 then left3 ‘Is sw3 left active? if porta.4 = 0 then right3 ‘Is sw3 right active? if porta.3 = 0 then left2 ‘Is sw4 left active? if porta.2 = 0 then right2 ‘Is sw4 right active? if porta.1 = 0 then left1 ‘Is sw5 left active? if porta.0 = 0 then right1 ‘Is sw5 right active? goto start ‘Routines for servomotor 1 left1: s1 = s1 + 1 if s1 = s2 then b0 = b0 + 1 ‘Increase the pulse widths1 = 0endifif b0 > 254 then max0 ‘Maximum 2.54 millisecondsgoto startright1:s1 = s1 + 1if s1 = s2 thenb0 = b0 1 ‘Decrease the pulse widths1 = 0endifif b0 < 75 then min0 ‘Minimum .75 millisecondgoto startmax0:b0 = 254 ‘Cap max b1 at 2.54 millisecondsgoto startmin0:b0 = 75 ‘Cap min b1 at .75 millisecondgoto start‘Routines for servomotor 2left2:s1 = s1 + 1if s1 = s2 thenb1 = b1 + 1 ‘Increase the pulse widths1 = 0endifif b1 > 254 then max1 ‘Maximum 2.54 millisecondsgoto startright2:s1 = s1 + 1if s1 = s2 thenb1 = b1 1 ‘Decrease the pulse widths1 = 0endifif b1 < 75 then min1 ‘Minimum .75 millisecond214 Chapter Twelve goto startmax1:b1 = 254 goto startmin1:b1 = 75 goto start‘Routines for servomotor 3left3:s1 = s1 + 1if s1 = s2 thenb2 = b2 + 1 s1 = 0endifif b2 > 254 then max2 goto startright3:s1 = s1 + 1if s1 = s2 thenb2 = b2 1 s1 = 0endifif b2 < 75 then min2 goto startmax2:b2 = 254 goto startmin2:b2 = 75 goto start‘Routines for servomotor 4left4:s1 = s1 + 1if s1 = s2 thenb3 = b3 + 1 s1 = 0endifif b3 > 254 then max3 goto startright4:s1 = s1 + 1if s1 = s2 thenb3 = b3 1 s1 = 0endifif b3 < 75 then min3 goto startmax3:‘Cap max b1 at 2.54 milliseconds‘Cap min b1 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecond‘Cap max b2 at 2.54 milliseconds‘Cap min b2 at .75 millisecond‘Increase the pulse width‘Maximum 2.54 milliseconds‘Decrease the pulse width‘Minimum .75 millisecondRobotic Arm 215 b3 = 254 ‘Cap max b3 at 2.54 milliseconds goto start min3: b3 = 75 ‘Cap min b3 at .75 millisecond goto start ‘Routines for servomotor 5 left5: s1 = s1 + 1 if s1 = s2 then b4 = b4 + 1 ‘Increase the pulse width s1 = 0 endif if b4 > 254 then max4 ‘Maximum 2.54 milliseconds goto start right5: s1 = s1 + 1 if s1 = s2 then b4 = b4 1 ‘Decrease the pulse width s1 = 0 endif if b4 < 75 then min4 ‘Minimum .75 millisecond goto start max4: b4 = 254 ‘Cap max b4 at 2.54 milliseconds goto start min4: b4 = 75 ‘Cap min b4 at .75 millisecond goto start end In the above program variable S2 is assigned a value of 4. To increase the speed of the servomotor’ s movement, decrease this value . To slow down the servomotor movement, increase this value. Increasing the Lifting Capacity of the Robotic Arm Substituting the top two HS322 servomotors connected to the gripper with two HS85MG servomotors can increase the lifting capacity of the robotic arm. The HS85MG servomotors are substantially smaller and lighter , while producing close to the same torque as the HS322 servomotors. The downside is that the HS85MG servomotors cost about 3 times the amount of the HS322 servomotors. Do not try to substitute the HS85BB servomotor for the HS85MG. The HS85BB uses plastic gears , which will strip pretty quic kly. The HS85MG incorporates metal gears that last. To use the HS85MG servomotors in the robotic arm,substitute the top HS322 bracket for an HS85MG brac ket. In addition you need to order the ser vomotor gripper that has been modified to use an HS85MG servomotor .216 Chapter Twelve Adding a Robotic Arm Base The weakest link in the robotic arm, as it stands right now, is the base servomotor. The bearing in the bottom servomotor is subjected to all the stress and weight of the entire arm as it turns and lifts any object. We can greatly improve upon this situation by adding a second bearing that removes most of the stress on the servomotor’s small bearing. To incorporate this second bearing, we need to build a small base . I tried a number of designs. The one that I feel works best is made primarily from 3/4inthick hardwood. The following drawings show the five pieces needed to make the base . Figures 12.35 and 12.36 show the wood bloc ks needed for mounting the base servomotor. Figures 12.37 and 12.38 show the sides for the base. Figure 12.39 is a metal baseplate. The two servomotor blocks are mounted to the baseplate , using wood screws through the bottom. The servomotor is mounted to the wood bloc ks (see Fig. 12.40). Next the side pieces are mounted to the wood bloc k (see Fig. 12.41).We need a 0.40in, “length of 1”indiameter wood dowel. To this piece of wood we center and attac h a round servomotor horn, using two small wood screws (see F ig. 12.42). The top of the servomotor horn should be sanded flat to remove the small lip around the center . The wood dowel is fitted onto the base servomotor (see Fig. 12.43). Next the 3insquare bearing is placed on the sides to ensure everything lines up properly. The wood dowel should be centered in the bearing (see Fig. 12.44). Mount the bearing to the sides, using four wood screws. A top plate for the 3insquare bearing is shown in Fig. 12.45. This plate is mounted to the bearing using four 632 plastic mac hine screws and nut. 1.0 1.5 Material 3/4 -thick hardwood All holes 1/16 diameter Semicircle 3/8 diameter .158 .555 .945 Top C/L Bottom .375 .375 1.125 All dimensions in inches Figure 12.35 Servomotor block A.1.0 1.5 Material 3/4 -thick hardwood All holes 1/16 diameter .158 .555 .945 Top Side C/L Bottom .375 .375 1.125 All dimensions in inches Figure 12.36 Servomotor block B. 3.5 1.7 Material 3/4 -thick hardwood Side Top Bottom All holes 1/16 diameter .216 .375 .228 2.8 .5 3.0 All dimensions in inches Figure 12.37 Side block A. 2173.5 1.7 Material 3/4 -thick hardwood Side Top Bottom All holes 1/16 diameter .534 .375 .228 2.8 .5 3.0 All dimensions in inches Figure 12.38 Side block B. 1/8 -3/16 aluminum or CRS .5 3.0 .375 2.625 .696 3.119 1.129 1.879 All holes 3/16 diameter, countersunk on bottom. All dimensions in inches. Figure 12.39 Baseplate. 218Robotic Arm 219Figure 12.40 Assembling servomotor blocks and servomotor to base. Figure 12.41 Attaching sides to base.220 Chapter Twelve Figure 12.42 A 1in 3 0.40in wood dowel with round servomotor horn. Figure 12.43 Attaching a servomotor horn to servomotor base . When the top plate is secured to the bearing , the top of the wood dowel should be right underneath the top plate. Place the bottom servomotor bracket of the robotic arm on top of the top plate . Secure the servomotor brac ket (and top plate) to the underlying dowel through the four center holes in the top bearing plate (see Fig. 12.46). The top section of the robotic arm is fitted to the base servomotor bracket. The finished robotic arm is shown in F igs. 12.47 and 12.48. In the picture note the use of the smaller HS85MG servomotors connected to the gripper.Robotic Arm 221Figure 12.44 Attach 3insquare bearing to base, check for alignment. 1.69 1.69 1.32 1.32 .219 .219 Corner holes 5/32 dia. Center holes 1/8 dia. Material: Aluminum 3 3 3 3 .042 thick All dimensions in inches Figure 12.45 Top bearing plate.Figure 12.46 Attach top bearing plate, servomotor bracket to 3insquare bearing. Figure 12.47 Finished robotic arm with base (right side). 222Robotic Arm 223 Figure 12.48 Finished robotic arm with base (left side). Parts List Robotic arm (3) HiTec servomotors (HS322HD) (2) HiTec servomotors (HS475HB) (5) Servomotor bracket assemblies (1) Servomotor gripper assembly (1) Base board (5) 12or 24in servomotor extensions Base Servomotor blocks A and B Baseplate Base sides A and B224 Chapter Twelve 3insquare bearing Top bearing plate 1indiameter 3 0.40in wood dowel Plastic machine screws and nuts, wood screws Available from Images SI Inc. (see Suppliers at end of book). Fourservomotor controller (1) PIC 16F84 (1) 4MHz Xtal (2) 22pF capacitors (4) SPDT PCmounted switches with centeroff position (8) 10kV, 1/W resistors 4 (1) 4.7kV, 1/W resistor 4 (1) 0.1mF, 50V capacitor 5V power supply (regulated) Kit available from Images SI Inc. (see Suppliers). Fiveservomotor controller (1) PIC 16F873 (1) 4MHz Xtal (2) 22pF capacitors (5) SPDT PCmounted switches with centeroff position (11) 10kV, 1/W resistors 4 (1) 4.7kV resistor (1) 0.1mF, 50V capacitor 5V dc power supply Kit available from Images SI Inc. (see Suppliers).Chapter 13 Bipedal Walker Robot In this chapter we will construct and program a bipedal walking robot (see Fig. 13.1). Bipedal robots more closely resemble humans because they use two legs to walk. Bipedalism is a necessary step to creating advanced robots that can work and function in human environments . The heart and mind of this robot are the 16F84 microcontroller. The microcontroller will be programmed using the PicBasic (or PicBasic Pro) compiler . Muscle for motion is generated using a series of eight HS322 servomotors, four servomotors for each leg. I have not taken any shortcuts in building this bipedal robot, meaning this robot is a true bipedal walker robot. This criterion demands that the robot balance itself on one leg in order to lift the other leg to initiate w alking. This action is accomplished using independent ankle , knee, and hip movements . This bipedal robot does not have oversized feet or footpads. This eliminates the type of lowtec hnology tilting bipedal w alker that uses oversized feet to keep the robot from tipping over when movement proceeds from one leg to the oth er. You may have seen this type of “bigfoot” walker; the older units ha ve a motoractivated cam that rises and moves one leg after the other . Lately I’ve seen servomotorpowered tilting bigfoots on the loose . To see a movie of this bipedal robot w alking, go to the Internet to the following page: www.imagesco.com/catalog/biped/walker.html. My design calls for using four servomotors in each leg (see Fig. 13.1). The initial walking gait programmed into the robot resembled that of the flamingo bird. This particular bird has a reverse knee joint. If that bird doesn’t bring a clear enough picture to mind, perhaps the Imperial w alker from the original Star Wars film(s) will suffice. Nature has evolved a threejointed leg for most w alking animals. Although it may appear that our robotic leg has four joints , because it has four servomotors, it is essentially a threejointed leg. The reason is that our first and second servomotors, starting from the bottom of the leg , form a twodirectional ankle. It is important that the ankle can tilt the foot, left to right as well as Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 225226 Chapter Thirteen Figure 13.1 Bipedal robot ready to walk. forward and backward. Humans have twodirectional ankles; this requires two servomotors to replicate in our leg . So the third servomotor is considered the knee joint, and the forth servomotor the hip joint. A Question of Balance? When we w alk, we receive constant feedbac k from our leg musc les and feet such as stretch, tension, and load, in addition to having tilt and balance information present from our inner ear. Remove this physical feedback information and remove any visual c lues, and it becomes muc h harder to w alk. Imagine how much harder, if not impossible, it would be to learn how to w alk without sensory feedback. This lack of feedbac k is a dilemma for robotics . It is possible to program a bipedal walker robot to walk without feedback and a sense of balance. To do so, exact position control and movements are measured for eac h leg servomotor action, each action sequence is programmed into the microcontroller , the program is initiated, and the sequence repeated to ac hieve a walking gait.Bipedal Walker Robot 227 Figure 13.2 FlexiForce pressure sensor. This bruteforce programming works, but it is not adaptive. If any weight on the robot shifts (battery pack moves) or if you have the robot carry a weight, anything that changes the robot’s center of gravity, then the program will need to be adjusted. A little sensory feedback may help the robot walk and be more adaptive. A Little Feedback Feedback comes in many forms. The sensor I would incorporate into this robot is a pressure sensor. I will be placing a pressure sensor on the base of each footpad. The sensor could tell the microcontroller when there is no pressure (weight) on a foot. This could be used to adaptively tilt the robot until there is no weight on the opposite footpad. The sensor is a FlexiF orce pressure sensor (see F ig. 13.2). (FlexiForce is a trademark of Tekscan, Inc.) This particular sensor is made to detect pressure from 0 to 1 lb. Although the final weight of the robot may be slightly more the sensor top weight, I feel it’ s a better (more sensitive) c hoice than taking the next sensor that measures pressure between 0 and 25 lb . The pressure sensor is a variableresistor type . As pressure increases , its resistance drops. Since we are using the sensor to determine when there is zero weight on a leg, we don’t need to perform an A/D conversion to read varying pressure (weight). Instead we can use an opamp and comparator. The opamp converts the resistance c hange in the sensor to an electric c hange. The comparator is set to trigger on zero weight. The output of the comparator can be read by the microcontroller as a simple highlow signal. This bipedal robot does not use any feedbac k, so it is not adaptive to shifting weight loads. I have provided this feedbac k information in case you wish to advance this basic bipedal w alker on your own. ServomotorsThis bipedal walker utilizes common inexpensive HiTec HS322HD 42oz torque servomotors. Other more powerful servomotors are a vailable, such as the HS425 and HS475, and they will increase the weightcarrying capacity of the robot. However, these more powerful servomotors also require greater electric current. So the battery pack will need to be increased proportionally. The robot, as it stands, is capable of carrying its own 6V battery pac k and circuitry.228 Chapter Thirteen Figure 13.3 Servomotor brackets needed for one leg. Servomotor Brackets This robot uses the same servomotor brackets as outlined in Chap. 12. That information will not be repeated here . In Fig. 13.3 the brackets needed for one robot ic leg are shown. You need two suc h sets of servomotor brac kets, eight in all, to build this bipedal robot. The servomotor horns used on these servomotor brac kets are included with all the compatible HiTec servomotors, such as HS322, HS425, HS475, and HS35645. These brackets may also be used with similar size Futaba servomotors, but you may have to purchase the horns separately. Footpads The footpads for the robot are shown in Figs. 13.4 and 13.5. I glued rubber gasket material to the bottom of the plastic footpad to make the pad nonskid. The footpads provide a larger surface area that makes it easier for the biped to balance and walk. They are attached to the bottom U bracket of the bottom servomotor. I arbitrarily chose to make the footpad size 1.5 in wide 3 4 in long. I cut out this size rectangle from 1/4inthick acrylic plastic. The location of the servomotor bracket on the feet is shown in Fig. 13.4. You will notice the bracket is not centered on the plastic foot;it is located at one side toward one end (considered the back). Drill four 1/indiameter holes in the 8 plastic that line up with the four holes on the U brac ket. Each drilled hole must be countersunk on the bottom of the foot, so that the mac hine screw head will not protrude from the bottom of the foot; see side view and closeup of Fig. 13.4 and finished footpad in F ig. 13.5. This will allow the foot to lie flat against the floor. On the prototype the corners of the footpads are square (see F ig. 13.5). I plan to round the corners of the footpads , so they will be less likely to catc h on something and trip the robot when walking. The footpads are attached to the U bracket using four 440 machine screws, nuts, and lockwashers.Countersunk hole (see text) Plastic Bracket Close-Up Outside Edge Top View: Left and Right Foot Material: 1/4-3 1.5-3 4.0-in Transparent Plastic Servomotor Bracket Placement Outside Edge Front 1.5 in 4.0 in Side View Figure 13.4 Diagram of footpad. Figure 13.5 Picture of footpad. 229230 Chapter Thirteen 2 1 1 C/L C/L Material 1/8 3 1 3 4 aluminum Hole size 5/32 dia. All dimensions in inches Figure 13.6 Aluminum hip bar. The bottom of the acrylic plastic feet can be slippery , depending upon the surface material the bipedal robot is walking on. I glued soft rubber sheet gasket material to the bottom of the acrylic feet to create a nonskid bottom surface for the feet. If just the front and bac k of the gasket material are glued to the plastic foot, a small flat pocket is created in the center section of the foot. This flat pocket is ideal for locating a flat sensor that could be slid in between the gasket material and the acrylic plastic. Although we will not be using any flat sensor in this robot, it could become a future modification, and you ma y want to leave this option open when gluing the gasket material to the footpad. I have found this robot biped walks and balances so easily that I believe it’s possible to reduce the size of the footpads or remove them entirely . This idea is open for future experimentation. The hip bar that connects the top servomotor brackets of both legs is shown in Fig. 13.6. The base material is 1/inthick aluminum bar 1 in wide 3 4 in 8 long. Mark a centerline (C/L) across the width and the length,as shown in Fig. 13.6. From the width C/L mark another line 1 in a way from the C/L on eac h side. Next use the base of the servomotor brac ket to mark the four mounting holes. Align the bracket on the left side so that an “X” from the drawn centerlines is centered in the rightmost hole. Mark the four holes with a pencil.Align the bracket on the right side so that an “X” from the drawn centerlines is centered in the leftmost hole. Mark the four holes with a pencil. Punch the center of eac h hole with a hammer and punc h. Drill the punc h holes with a 5/32in drill. Clean eac h hole to remove any burrs with a file or deburring tool. Assembly When you assemble the servomotors to the servomotor brac kets, center each servomotor before attaching the servomotor shaft to the hornbrac ket assemBipedal Walker Robot 231 Figure 13.7 Bipedal robot with all servomotors centered. bly. The walking program expects the servomotors to be aligned in this w ay. If a centering servomotor signal is sent to all eight servomotors , the robot walker will appear as shown in Fig. 13.7. This is not the start position of the w alking program. Schematic Figure 13.8 is the sc hematic of our bipedal w alker robot. To achieve maximum torque from the servomotors , I needed to run them at 6 V. To run the PIC 16F84 at close to 5 V, I incorporated a 1N4007 diode. The average voltage drop across a silicon diode is 0.7 V. So at peak power from the batteries (under load) the microcontroller will receive about 5.3 V, which is within the voltage range for this microcontroller. A photograph of the prototype circuit is shown in F ig. 13.9. The battery pack I used is below the circuit board. It holds four AA batteries. I used a small piece of Velcro to secure the battery pack to the hip bar . I secure the circuit board by using two small elastic bands (see Fig. 13.10).232 Chapter Thirteen RB7 RB6 RB5 RB4 RB3 RB2 RB1 RB0/INT RA4/TOCKI RA3 RA2 RA1 RA0 13 12 11 109876321 18 17 Servo Motor 0 MCLR ’ OSC1 OSC2 VDD VSS 5 416 15 14 U1 R1 4.7 KW X1 4.0 MHz +6 V PIC 16F84 Caps 22 pF To Servo # 7 To Servo # 6 To Servo # 5 To Servo # 4 To Servo # 3 To Servo # 2 To Servo # 1 To Servo # 0 +6 V Reset Switch Figure 13.8 Bipedal robot schematic. Figure 13.9 Top view of prototype circuit board. The four AA battery, 6V power supply only lasts a short time . The bipedal robot appears to be able to lift more weight than I placed on it,so you may be able to add a second 6V power supply and increase the untethered w alking time. In any case I only use the battery pac k for demonstrations. For most developmentBipedal Walker Robot 233 Figure 13.10 Side view of circuit board and battery pack attached to robot. work you may want to build an external regulated power supply for the biped, as I have, and tether the power supply to the robot. Keep the unused battery pac k on the robot, so you will not ha ve to compensate for the additional weight when demonstrating the robot’s walking ability using the battery pac k. Program When the robot is assembled, you may have to adjust the program slightly . There will be slight variances in your servomotor positions as compared to my prototype due to small variances in the construction. You only need to add or remove one line in the entire program to make adjustments , and the line is: goto holdThe hold subroutine keeps the servomotors loc ked in their last position. The robot stays frozen, giving you plenty of time to look over its position. This is the procedure for using that one line and adjusting the program.You place that line after eac h robotic movement. Check the position, adjust the movement if necessary, check again, and adjust if necessary until the position is perfect. Movement is adjusted by varying the Y1 and Y2 numbers in eac h movement. I cannot imagine the variance being more that ±5 points off what the program is showing. There are 15 movements to c heck. I would advise letting the robot step through eac h movement; you will see if there is a problem. The robot ma y either trip on its feet or lose its balance . If that happens, you know you ha ve to adjust that movement. But you must work it through movement by move 234 Chapter Thirteen Figure 13.11 Front view of robot. ment. If you just try to let the walker walk, it will be hard for you to determine which movement (if any) is causing a problem. The first thing to c heck is the start position of the robot. Write the goto hold line right after the command Gosub servoout . The robot should be level, standing in a position shown in F igs. 13.11 and 13.12. If adjustments are necessary, you need to make them in the “initialize variables” section. Once you are satisfied, remove the goto hold line you wrote in the program. Place the goto hold line at the end of the “First movement.” Check position, adjust if necessary, then move the goto hold line to the end of the “Second movement.” Continue in this manner until all movements have been checked. The way the program is written, the robot will take three steps and then stop. You can change the range of B(10) to increase or decrease the amount of steps taken. Subroutines M1, M2, and M3 The subroutines M1, M2, and M3 are dela y routines. These routines slow the servomotor movement, so the movement is smooth.Without these routines theBipedal Walker Robot 235 Figure 13.12 Side view of robot. servomotors would jerk into position so quic kly that the motion would topple the robot. The reason for three routines is that I want to affect two independent servomotor motions at the same time. The numbers controlling the servomotor positions could be both (1) decreasing (M1 –,–) and increasing (M2 1,1) and (2) increasing and decreasing (M3 1,–). Hence we need the three subroutines to handle the motion. ‘Bipedal walker program‘Declare variablesx1 var bytex2 var bytey1 var bytey2 var bytelp var byte‘Declare array236 Chapter Thirteen b var byte[12]‘Initalize array variablesb(0) = 148 ‘Right ankle (vertical) b(1) = 121 ‘Right ankle (horiz.) b(2) = 204 ‘Right knee b(3) = 126 ‘Right hip b(4) = 150 ‘Left ankle (vertical) b(5) = 178 ‘Left ankle (horiz.) b(6) = 101 ‘Left knee b(7) = 180 ‘Left hip b(8) = 0 ‘Counter b(9) = 0 ‘Counter b(10) = 0 ‘Counter b(11) = 0 ‘Dummy value start:‘Holding loop that holds upright position 3 seconds before movingb(8) = b(8) + 1gosub servooutif b(8) < 180 then goto startb(8) = 0 ‘Reset loop counter‘——————————————————for b(10) = 1 to 3 ‘Take 3 steps forward‘——————————————————‘Leg movements for one whole step‘——————————————————‘First movementx1 = 0 ‘Servomotor 0x2 = 4 ‘Servomotor 4y1 = 129 ‘Tilt right ankle (horiz.)y2 = 135 ‘Tilt left ankle (horiz.)lp = 106 ‘Loop countergosub m1‘——————————————————‘Second movementx1 = 5 ‘Servomotor 5x2 = 6 ‘Servomotor 6Bipedal Walker Robot 237 y1 = 2 ‘Left ankle (vert.) y2 = 70 ‘Left knee lp = 140 ‘Loop counter gosub m3 ‘—————————————————— ‘Third movement x1 = 4 ‘Servomotor 4 x2 = 7 ‘Servomotor 7 y1 = 150 ‘Left ankle y2 = 160 ‘Left hip lp = 75 ‘Loop counter gosub m3 ‘—————————————————— ‘Fourth movement x1 = 1 ‘Servomotor 1 x2 = 5 ‘Servomotor 5 y1 = 132 ‘Straighten ankle y2 = 200 ‘Straighten ankle lp = 90 ‘Loop counter gosub m3 ‘—————————————————— ‘Fifth movement x1 = 6 ‘Servomotor 6 x2 = 0 ‘Servomotor 0 y1 = 85 ‘Left knee y2 = 140 ‘Straighten ankle lp = 96 ‘Loop counter gosub m2 ‘—————————————————— ‘Sixth movement x1 = 0 ‘Servomotor 0 x2 = 4 ‘Servomotor 4 y1 = 167 ‘Tilt ankle left y2 = 169 ‘Tilt ankle left lp = 80 ‘Loop counter gosub m2 ‘—————————————————— ‘Seventh movement x1 = 6 ‘Servomotor 6 x2 = 5 ‘Servomotor 5 y1 = 101 ‘Tilt knee y2 = 178 ‘Tilt ankle lp = 95 ‘Loop counter gosub m3 ‘——————————————————238 Chapter Thirteen ‘Eighth movementx1 = 2 x2 = 1 y1 = 244 y2 = 104 lp = 140 gosub m3‘——————————————————‘Ninth movementx1 = 3 x2 = 7 y1 = 140 y2 = 180 lp = 80 gosub m2‘——————————————————‘Tenth movementx1 = 1 x2 = 2 y1 = 121 y2 = 204 lp = 150 gosub m3‘——————————————————‘Eleventh movementx1 = 0 x2 = 4 y1 = 129 y2 = 135 lp = 150 gosub m1‘——————————————————‘Twelfth movementx1 = 5 x2 = 6 y1 = 217 y2 = 70 lp = 144 gosub m3‘——————————————————‘Thirteenth movementx1 = 4 x2 = 3 y1 = 133 y2 = 126 ‘Servomotor 2‘Servomotor 1‘Right ankle‘Right knee‘Loop counter‘Servomotor 3‘Servomotor 7‘Right hip‘Left hip‘Loop counter‘Servomotor 1‘Servomotor 2‘Right ankle‘Right knee‘Loop counter‘Servomotor 0‘Servomotor 4‘Straighten right ankle‘Straighten left ankle‘Loop counter‘Servomotor 5‘Servomotor 6‘Left ankle‘Left knee‘Loop counter‘Servomotor 4‘Servomotor 3‘Left ankle‘Right hipBipedal Walker Robot 239 lp = 66 ‘Loop countergosub m1‘——————————————————‘Fourteenth movementx1 = 6 ‘Servomotor 6x2 = 5 ‘Servomotor 5y1 = 101 ‘Left kneey2 = 178 ‘Left anklelp = 144 ‘Loop countergosub m3‘——————————————————‘Fifteenth movementx1 = 0 ‘Servomotor 0x2 = 4 ‘Servomotor 4y1 = 148 ‘Left kneey2 = 150 ‘Left anklelp = 115 ‘Loop countergosub m2‘——————————————————next b(10) ‘Next step‘——————————————————hold: ‘Hold positiongosub servooutgoto hold‘——————————————————servoout:‘Output servomotor position(s)portb = 0 ‘Prevent signal inversion‘Right legpulsout portb.0, b(0) pulsout portb.1, b(1) pulsout portb.2, b(2) pulsout portb.3, b(3) ‘Send current servo 1 position out‘Send current servo 2 position out‘Send current servo 3 position out‘Send current servo 4 position out‘Left legpulsout portb.4, b(4) pulsout portb.5, b(5) pulsout portb.6, b(6) pulsout portb.7, b(7) ‘Send current servo 5 position out‘Send current servo 6 position out‘Send current servo 7 position out‘Send current servo 8 position outpause 5 ‘5 millisecond delay to generate 50 Hz signalreturn ‘To servomotors‘——————————————————240 Chapter Thirteen m1: b(8) = b(8) + 1if b(9) = 2 then m12b(9) = b(9) + 1goto m13m12:b(x1) = b(x1) 1b(x2) = b(x2) 1b(9) = 0m13:if b(x1) < y1 thenb(x1) = y1endifif b(x2) < y2 thenb(x2) = y2endifgosub servooutif b(8) < lp then m1b(x1) = y1b(x2) = y2b(8) = 0b(9) = 0return‘——————————————————m2: b(8) = b(8) + 1if b(9) = 2 then m22b(9) = b(9) + 1goto m23m22:b(x1) = b(x1) + 1b(x2) = b(x2) + 1b(9) = 0m23:if b(x1) > y1 thenb(x1) = y1endifif b(x2) > y2 thenb(x2) = y2endifgosub servooutif b(8) < lp then m2b(x1) = y1b(x2) = y2b(8) = 0b(9) = 0‘(negative increment(s) ,)‘(positive increment(s) +,+)Bipedal Walker Robot 241 return‘——————————————————m3: ‘(positive negative increment +,)b(8) = b(8) + 1if b(9) = 2 then m32b(9) = b(9) + 1goto m33m32:b(x1) = b(x1) + 1b(x2) = b(x2) 1b(9) = 0m33:if b(x1) > y1 thenb(x1) = y1endifif b(x2) < y2 thenb(x2) = y2endifgosub servooutif b(8) < lp then m3b(x1) = y1b(x2) = y2b(8) = 0b(9) = 0returnGoing Further There are many areas for improvement. One of the simplest tasks you can perform is to reduce the loop counter (LP) variable in eac h movement. I exaggerated this number to ensure that the servomotors got to their proper position. The walking gait used in this robot w as the first one I developed. I am sure there is muc h room for improvement for anyone who w ants to take the time and develop one. In addition, you can try to program completely different walking gaits. Right now the robot used two reverse knee joints . I looked at the robot stance using one reversed knee and one forward knee. It appears to have better been balanced than the current two reverse knee biped stance . In the future I ma y try to develop a gait using a forw ard and reverse knee stance . This would most definitely be a robotic gait, since I don’t believe there is any animal that uses both a reverse and a forward knee leg for locomotion. This is another area you may want to work on.242 Chapter Thirteen Turning right and left As the bipedal w alker stands, it can only w alk forward. While I was developing the w alking program, I happened across an interesting accident. On certain occasions the robot would pivot to the left or to the right. I plan on developing this “accident” to see if I can use it to turn the robot to the left and right. If you want to attempt this, here are the basic instructions. To make the robot pivot, first raise one leg . On the raised leg tilt the horizontal ankle servomotor slightly, and then tilt the vertical ankle servomotor up slightly . Next place the weight back down on the raised leg;the robot will pivot as the weight shifts. Turning the robot in this manner must be accomplished incrementally . Try to turn it too muc h at one time, and the robot will topple. Finally with a little work, you should be able to make the robot w alk backward. I mentioned adaptive walking and balance control.We used all eight pins of port B on the PIC 16F84,but the PIC 16F84 still has five unused pins assigned to port A. These pins could be used for programming options suc h as adaptive balance control, or a run/walk switch, perhaps a forward/backward switch, or even a turn left/right sensor. So you see , we have only scratc hed the surface of pla ying with this biped walker. There is much one can do and learn from this project. Parts List (8) HiTec servomotors (or similarsize and torque servomotor) (HS322) (8) Servomotor brackets (2) 1/in 3 1.5in 3 4in acrylic plastic 4 (1) PIC 16F84 (4MHz) (1) 4.0MHz Xtal (2) 22pF capacitors (1) Diode (1N4007) (1) 6V AA battery holder (flat) (8) Threeposition headers (for connecting servomotors to circuit) Plastic screws and nuts, Velcro, prototyping breadboard, elastic bandsChapter 14 Color Robotic Vision System The robot we will build in this chapter will be capable of looking at and seeing an object (or target) and following that object. If the object gets too close to the robot, the robot will bac k away from it. Choose a bright target that will ha ve good contrast with the bac kground environment. Colored objects are fine; a bright red or yellow ball works fine . For my tests I used a 2.5in square of orange construction paper taped to a stiff wire . One of the most difficult areas in robotics toda y is the creation of an artifi cial vision system for a robot to see. Teaching a robot to see is not simply a case of connecting a video camera to a computer. The electronic representation of an image created by a video camera must be presented in a w ay a computer can “look at” and interpret (“see”) the image. To gain an understanding of the processes involved, let’s examine how a computer might look at a simple blac kandwhite image.We must first define the resolution of the picture image in pixels. For our discussion let’s assume a lowresolution picture of 80 3 143 pixels. At that resolution our computer must look at 11,440 pixels (80 3 143 5 11,440). Each pixel of the picture can be any tonality of gray between pure black and pure white.We now have to determine how many different shades our computer can differentiate between pure black and pure white. If we wanted each pixel to be represented by a single byte (8bit number), there would be 256 shades of gra y, including pure white (0) and black (255). The computer would look at each pixel and assign a number between 0 and 255 depending upon its tonality (grayness). After the computer assigned numbers to eac h 11,440 pixels , it transformed the basic image into a numbered representation it needs to “look” at the image. The software that looks at an image and interprets visual features is appropriately called image processing. Robotists over the years ha ve gleaned some techniques for helping computers to see . One technique is called edge detection. Here the computer looks through the image . You program the computer Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 243244 Chapter Fourteen Figure 14.1 Field of view from CMU camera. so that any pixel dark enough to be a 200 number or larger is a possible edge , and you c hange that pixel to pure blac k, number 256. Any number less than 200 probably isn’t an edge , and the computer c hanges those pixels to pure white, number 1. What is left is a simplified representation of the picture that can be more easily analyzed. The same process used for edge detection ma y also be used to detect particular colors in an image or the contrast between colors. Once an object has been detected, through contrast, color, or edge detection, the processing software can assign location parameters to the object within the image and field of view (FOV) of the camera. In Fig. 14.1 we have a representation of the FOV from the CMU camera we will be using and a few of the image processing parameters available. Once an object is detected by the camera, we can read these image processing parameters in real time from the serial communication port of the camera. We use these parameters to trac k an object in the camera’ s image space and to move our robot accordingly. CMU CameraThe CMU camera (see Fig. 14.2) was developed at Carnegie Mellon University (CMU). The CMU camera uses an SX28 microcontroller interfaced to an Omnivision OV6620 CMOS camera chip. The SX28 microcontroller does much of the image processing for us . We communicate with the camera via a stan dard RS232 or TTL serial port. A few of the CMU camera features are as follows: Tracks userdefined color objects at 17 frames per second Finds the center of the objectColor Robotic Vision System 245 Figure 14.2 Front CMU camera. Gathers mean color and variance data Resolution of 80 3 143 pixels Serial communication at 115,200, 38,400, 19,200, and 9600 Bd Demo mode that automatically loc ks onto and drives a servomotor to trac k an object Serial Communication As stated, we communicate to the CMU camera via a serial interface .We will create a serial communication link between the CMU camera and both a personal computer (PC) and the PIC microcontroller .We will first look at the PC communication to the CMU camera. Figure 14.3 is a simple Windows 98 program. It allows you to test the CMU camera and the serial communication link (port number and baud rate). You can adjust the PC’s baud rate and serial port through dropdown menu items . This program ma y be downloaded without cost from this website: http://www.cmucam.com. Before you start the Windows program, you need to set up the CMU cam era’s baud rate . Figure 14.4 shows the bac k of the CMU camera, where the male header is located to place various jumpers . The baud rate is selected using jumper 2 and jumper 3 on the bac k of the CMU camera:246 Chapter Fourteen Figure 14.3 Basic serial Windows PC communication program. Figure 14.4 Back of CMU camera showing baud rate jumpers .Color Robotic Vision System 247 Baud rate Jumper 2 Jumper 3 115,200 Open Open 38,400 Set Open 19,200 Open Set 9,600 Set Set For PC communication I recommend using the 115,200Bd rate . Use this baud rate because once you have your simple communication up and running, you can switch over to a more sophisticated Windows program to evaluate the CMU camera parameters. With the baud rate set, connect the serial cable to the CMU camera (see Fig. 14.5). Connect a DB9 pin serial cable from the PC to the camera. Start the program. Set the program’s baud rate to match the CMU camera’s baud rate. Set the serial port to the one you connected to the CMU camera. If your computer has multiple serial ports, you may have to try different COMM ports to find out which one is connected to the camera. To test a port, set the serial port to COMM1. Turn on the CMU camera. The following message should be displa yed when the camera is turned on: cmucam V1.12:Figure 14.5 PC serial cable connection to CMU camera.248 Chapter Fourteen If you do not see this message , turn off the camera, set the serial port to COMM2, and test again. Continue in this manner until you find the right COMM port. If you don’t see the message with any of the COMM ports on your computer, you may have the baud rate set improperly—doublec heck. Once you see the message, you begin to communicate with the CMU camera. To turn on the camera’s green LED, enter the command l1 1and hit Return. To turn off the green LED, enter the command l1 2 and hit Return. Once you ha ve the communication link working , you are finished with the first program, and it is time to move onto the main VB application program. VB Application Program The VB application program is included on the CDROM with the CMU camera.The application allows you to see how the CMU camera images different scenes or targets. The current VB application isn’t stable; however, by the time this book goes to press, a newer, (hopefully) more stable application program will be a vailable. The simple application we used before provided the correct port number that you will need to allow this program to function properly. The baud rate used on this program is fixed at 115,200. So make sure the CMU camera is set at 115,200 Bd. Figure 14.6 Windows PC program.Color Robotic Vision System 249 Figure 14.7 Windows PC program showing frame dump. To view the image properly from the camera, hit the 180° option (see F ig. 14.6). Select the proper COMM port number and open the port. Turn on the camera. You should receive the “CMUcam V1.12” message. Hit the Dump Frame button and wait. It can take 10 s for the softw are to dump the frame. The image shown in the Dump Frame window (see Fig. 14.7) is a simple target I constructed. This target helped me calibrate the camera’ s field of view . The target is a 2.5in square of orange paper (see Fig. 14.8), held at a distance of 12 in from the camera lens . I also used this target to read the image processing parameters from my PIC program 2 as I moved the target left, right, up, and down. I assembled these image process readings in a small table;more about this later. You should use this opportunity to find a good target. Place the object you want to use as your target in front of the camera, and do a frame dump . You are looking to see that the object shows well in the image and has good contrast with the bac kground. You can also see how muc h space the object takes up in the image . This will give you an idea of how c lose you should hold the object to the camera. Once you have your target, you can start using the communication port for issuing commands to the CMU camera. Try turning the green LED on and off250 Chapter Fourteen Figure 14.8 Target used for calibrating CMU camera. as before. You can use this communication port to implement more c hallenging commands and see the results on the screen dump . Here are a few commands you ma y want to try: Turn on auto light adjustment. This command tells the camera to adjust to the ambient lighting . When you use this command, do not have your object/target in front of the camera. The command is cr 18 44. Now press the Return key or Send button. Wait 10 to 20 s for the camera to complete its ambient light adjustment. Then enter this command to turn off auto light adjustment: cr 18 44 19 32. Now press the Return key or Send button.This next command I found particularly useful. It turns on a fluorescent band filter with the auto light adjustment: cr 45 7 18 44. Now press the Return key or Send button. You can find other commands in the CMU manual. Interfacing the CMU Camera to a Robot The first step in interfacing the camera to a robot is to establish communication between the PIC microcontroller and the CMU camera. Remove the DB9 serial cable used for communicating with the PC . The camera has a TTL serial output next to the jumpers (see F ig. 14.9). Before we can use the TTL serial input/output pins , first we remove the MAX232 IC from the bac k of the CMU camera.Color Robotic Vision System 251 Figure 14.9 Back of CMU showing TTL serial communication jumpers.Note: At any time you need or w ant to reconnect the CMU camera serial interface to a PC , you will need to place the MAX232 c hip bac k onto the board. Plug the TTL cable onto the appropriate header pins on the CMU camera. Figure 14.10 is the schematic we will be using. You do not need to connect the two servomotors for our first two programs . PIC 16F84 Runs at 16 MHz One important note about the CMU schematic you must be aware of. The PIC 16F84 used in this circuit is a 20MHz version operating at 16 MHz with a 16MHz crystal. I needed to jump up in speed because the 9600Bd communica tion is running at the limit of the capacity of 16F84 at 4 MHz.To keep the baud rate timing accurate when we c hange clock speeds, we enter the command define osc 16This informs the compiler that we are running at 16 MHz. The compiler automatically adjusts the serial commands to keep the baud rate accurate . Program 1 This first program establishes a communication link between the CMU camera and PIC 16F84 microcontroller. It turns the green LED on the CMU camera on and off . You should not proceed to the more advanced programs until you have this program functioning properly.Figure 14.10 Main robot schematic. 252Color Robotic Vision System 253 When the program starts, it begins with a 5s countdown.If you look into the countdown loop, you will see that the program issues a reset command eac h time through the loop. I have found it necessary to send a few reset commands before the camera communication link becomes responsive . ‘PIC to CMU test‘Send serial information to CMU camera truedefine osc 16x var bytey var byterecdata var byte[10]trisb = 0portb = 0pause 1500serout portb.1,6,[“CMU Program V1”]for x = 0 to 4y = 5 xportb.3 = 1serout portb.1,6,[254,192,“Starting in”,#y]serout portb.2,2,[“RS”,13]pause 500portb.3 = 0pause 500next xserout portb.1,6,[254,1,“Resetting Cam.”]serout portb.1,6,[254,192] ‘Move to second lineserout portb.2,2,[“RS”,13]gosub displaypause 1000start:‘Turn green CMU LED onserout portb.1,6,[254,1,“Green LED On”]serout portb.1,6,[254,192] ‘Move to second lineserout portb.2,2,[“L1 1”,13]gosub displaypause 1000‘Turn green CMU LED offserout portb.1,6,[254,1,“Green LED Off”]serout portb.1,6,[254,192] ‘Move to second lineserout portb.2,2,[“L1 2”, 13]gosub displaypause 1000254 Chapter Fourteen goto startdisplay:serin2 portb.0,84,20,error,[str recdata\4]for x = 0 to 4serout2 portb.1,16468,[“ ”,#recdata[x]]recdata[x] = 32next xpause 1000returnerror:‘No acknowledgmentserout portb.1,6,[“No ACK Cont.”]pause 1000returnProgram 2 This second program displays on the LCD the major image processing param eters available from the CMU camera. This program just fits into the 1K memory space of the PIC 16F84. If you add a programming line or a couple of letters or spaces in any of the LCD displa ys, the program will not compile , because it will exceed the PIC 16F84 memory limit. Keep that in mind, if you encounter an error, when compiling this program. Incandescent or fluorescent lighting When I first starting working with the CMU camera, I was working under fluorescent lighting. The camera was not tracking its target as well as I expected. Going through the literature I had on the camera, I found a fluorescent filter. I incorporated the filter into my program, and the camera started trac king better. The program uses the fluorescent filter; it is in the following line: ‘Turn on fluorescent band filter and auto lighting adjustserout portb.2,2,[“CR 45 7 18 44”, 13]If you are using fluorescent lighting, you can leave this line alone. However, if your lighting is incandescent, change the command line to serout portb.2,2,[“CR 18 44”, 13]Obviously this program is more sophisticated than our first program. It displays the type S data pac ket and then displa ys the type M data pac ket in a loop for realtime object tracking. Let’s first look at the information that is provided in the type S data pac ket.Color Robotic Vision System 255 Type S Data Packet Displayed program parameter Item Description RM GM BM Rdev Gdev Bdev Rmean Gmean Bmean Rdeviation Gdeviation Bdeviation The mean red found in the current window The mean green found in the current window The mean blue found in the current window The deviation of red found in the current window The deviation of green found in the current window The deviation of blue found in the current window Here’s a listing of the information that is provided in the type M data pac ket. Type M Data Packet Displayed program parameter Item Description MMX mx The middle of mass x value MMY my The middle of mass y value LCX x1 The leftmost corner’s x value LCY y1 The leftmost corner’s y value RCX x2 The rightmost corner’s x value RCY y2 The rightmost corner’s y value pix pixel Number of pixels in the trac ked region conf confidence Number of pixels in area—capped at 255 It’s time to choose an object/target if you haven’t done so already. Program 2 needs an object to loc k onto, too. When the microcontroller runs , it displays information on the LCD screen. During the 10s autoadjust period, the camera should just be looking at the bac kground. When LED 1 (see sc hematic) starts to blink, place your object/target in front of the CMU camera. ‘CMU parameter display program‘By J. Iovinedefine osc 16recdata var byte[10]x var bytetrisb = 0portb = 0256 Chapter Fourteen pause 1500serout portb.2,2,[“RS”, 13]serout portb.1,6,[“CMU Test Program”]pause 1000serout portb.2,2,[“RS”, 13]serout portb.1,6,[254,1]‘Reset CMU cameraserout portb.2,2,[“RS”, 13]gosub display‘Turn green CMU LED onserout portb.2,2,[“L1 1”,13]gosub displayportb.3 = 1‘Turn on fluorescent band filter & auto lighting adjustserout portb.2,2,[“CR 45 7 18 44”, 13]gosub displayserout portb.1,6,[“Auto Adj.”]pause 10000 ‘Hold 10 secondsserout portb.1,6,[254,1]pause 50‘Turn off auto lighting adjustserout portb.2,2,[“CR 18 44 19 32”, 13]gosub display‘Turn green CMU LED offserout portb.2,2,[“L1 2”,13]gosub displayportb.3 = 0For x = 0 to 10 ‘Blink red LED to tell user to ready targetportb.3 = 1pause 250portb.3 = 0pause 250next x‘Set poll mode 1 packetserout portb.2,2,[“PM 1”, 13]pause 100‘Set raw dataserout portb.2,2,[“RM 3”, 13]pause 100Color Robotic Vision System 257 ‘Track window command looks at center of CMU window‘Grabs data and sends them to track color function‘Track:serout portb.2,2,[“TW”, 13]‘Gather the s statistics packet from TW commandserin2 portb.0,84,[str recdata\8]‘Display data on LCD screenserout portb.1,6,[“RM”,#recdata[2]]gosub holdserout portb.1,6,[“GM”,#recdata[3]]gosub holdserout portb.1,6,[“BM”,#recdata[4]]gosub holdserout portb.1,6,[“RDev.”,#recdata[5]]gosub holdserout portb.1,6,[“GDev.”,#recdata[6]]gosub holdserout portb.1,6,[“BDev.”,#recdata[7]]gosub holdpause 2000main:‘Send command track color (with no arguments)‘Will track last color grabbed by TW commandserout portb.2,2,[“TC”, 13]‘Gather the m statistics packet from TW commandserin2 portb.0,84,[str recdata\10]‘Display data on LCD screenserout portb.1,6,[“MMX”,#recdata[2]]gosub holdserout portb.1,6,[“MMY”,#recdata[3]]gosub holdserout portb.1,6,[“LCX”,#recdata[4]]gosub holdserout portb.1,6,[“LCY”,#recdata[5]]gosub hold258 Chapter Fourteen serout portb.1,6,[“RCX”,#recdata[6]]gosub holdserout portb.1,6,[“RCY”,#recdata[7]]gosub holdserout portb.1,6,[“Pix”,#recdata[8]]gosub holdserout portb.1,6,[“Conf”,#recdata[9]]gosub holdgoto main:display:serin2 portb.0,84,20,main,[str recdata\3]for x = 0 to 3serout2 portb.1,16468,[“ ”,recdata[x]]next xhold:pause 500serout2 portb.1,16468,[254,1]pause 40returnWhen the object is captured, the program first displa ys the S data pac ket. Then it goes into the main program loop, capturing and displaying the M data packet. Using this program, I constructed a data table that shows how my camera trac ked my object/target. In the following table , n/c 5 no c hange. Although this isn’t 100 percent accurate, I ignored small changes of less than a few points in either direction. The reasons are that (1) I don’t w ant anyone getting bogged down focusing on small c hanges and missing the important main changes and (2) when I moved a target to the left or right, I didn’t keep the height exactly in line . I just moved the target over and kept the height approximately the same. Obviously this caused minor fluctuations that can be ignored. Data Table ±X (Left and Right) Target Target Target Target Target Parameter 2 in left 1 in left centered 1 in right 2 in right MMX 67 57 45 35 20 MMY n/c n/c 74 n/c n/c LCX 53 45 33 23 4Color Robotic Vision System 259 LCY n/c n/c 47 n/c n/c RCX 80 70 58 53 28 RCY n/c n/c 102 n/c n/c PIX 144 150 163 162 162 CONF 31 215 232 142 39 From the above table we can make a general observation: As the target moves from left to right, MMX, LCX, and RCX decrease. The reverse is also true; as the target moves to the left, MMX, LCX, and RCX increase. Data Table ±Y (Up and Down) Target Target Target Target Target Parameter 2 in up 1 in up centered 1 in down 2 in down MMX n/c n/c 45 n/c n/c MMY 31 52 74 98 116 LCX n/c n/c 33 n/c n/c LCY 7 26 47 71 96 RCX n/c n/c 58 n/c n/c RCY 57 80 102 125 143 PIX 153 157 163 164 116 CONF 246 233 232 237 181 From the above table we can make a general observation. As the target moves up, MMY, LCY, and RCY decrease. The reverse is also true . As the object/target moves down, MMY, LCY, and RCY increase. Servomotors for robot This robot we will build uses two HS425 servomotors modified for continuous rotation. The procedure for modifying these servomotors for continuous rotation was discussed in Chap . 8. Once you ha ve the modified servomotors , it is essential that you determine the pulse widths needed for slow forw ard, slow backward, and stop. Figure 14.11 is a sc hematic for a circuit you can use along with the following PicBasic Pro program to determine the pulse widths . The pulse width is shown in real time on the LCD displa y. You c hange the pulse widths up or down by using the SPDT switch. It is essential that the switch used in this circuit have a centeroff position. ‘Continuous rotation servomotor calibration‘Serial communication to LCD display is 2400 baud invertedx var byte260 Chapter Fourteen Figure 14.1 1 Servomotor sc hematic for determining pulse widths for slow forw ard, slow backward, and stop. y var bytepause 1500serout portb.1,4,[“Servomotor Test”]pause 1000serout portb.1,4,[254,1]pause 20x = 150main:pulsout portb.0,xif porta.1 = 0 thenx = x + 1endifif porta.0 = 0 thenColor Robotic Vision System 261 x = x 1endifserout portb.0,4,[254,1,“ ”,#x]goto mainHere are the pulse width numbers I needed for the servomotors I used in my prototype robot. Function Right servomotor Left servomotor Stop 167 169 Slow backward 160 176 Slow forward 174 162 Note that the numbers represent 10ms increments in time.So the 167 used in the program is equal to 1.67 ms.Program 3 The following program is for our trac king robot. It uses information from the ±X data table to track an object/target from left to right. The PIX pixel parameter is used to determine range of the object. If the object (PIX gets too large) comes too close to the robot, the robot will back away from the object. Everything stated about program 2 also applies to this program.Keep in mind the lighting—fluorescent or incandescent—and remember to keep the target out of the camera’s FOV when it is adjusting for the ambient light. Again this pro gram just fits into the PIC 16F84; so if you add anything to the program, even a few spaces in the display, you stand a good chance of its not compiling properly. ‘CMU tracking program‘By J. Iovinedefine osc 16recdata var byte[10]x var byteconfid var bytetrisb = 0portb = 0pause 1500serout port.1,6,[“CMU Prg.”]serout portb.2,2,[“RS”, 13]pause 1250‘Reset CMU cameraserout portb.2,2,[“RS”, 13]262 Chapter Fourteen gosub display‘Turn green CMU LED onserout portb.2,2,[“L1 1”,13]gosub displayportb.3 = 1‘Turn on auto lighting adjust & fluorescent band filter ***serout portb.2,2,[“CR 45 7 18 44”, 13]gosub displayserout portb.1,6,[“A L”] ‘Auto lighting adjustmentpause 20000 ‘Hold 20 seconds‘Turn off auto lighting adjustserout portb.2,2,[“CR 18 44 19 32”, 13]gosub display‘Turn green CMU LED offserout portb.2,2,[“L1 2”,13]gosub displayportb.3 = 0‘Set poll mode1 packetserout portb.2,2,[“PM 1”, 13]pause 100‘Set raw dataserout portb.2,2,[“RM 3”, 13]for x = 0 to 10 ‘Blink red LED to tell user to ready targetportb.3 = 1pause 250portb.3 = 0pause 250next xportb.6 = 1 ‘Track LED on‘Track window command looks at center of CMU window‘Grabs data and sends it to track color function‘Track:serout portb.2,2,[“TW”, 13]pause 2000portb.6 = 0 ‘Track LED offmain:Color Robotic Vision System 263 portb.3 = 1‘Send commandtrack color (with no arguments)‘Will track last color grabbed by TW commandserout portb.2,2,[“TC”, 13]‘Gather the m statistics packet from TW commandserin2 portb.0,84,[str recdata\10]confid = recdata[9]if recdata[2] > 50 and confid > 20 then left ‘MMX if recdata[2] < 40 and confid > 20 then right ‘MMX if recdata[8] < 175 and confid > 25 then fwd ‘PIX if recdata[8] > 200 and confid > 25 then bwd ‘PIX serout portb.1,6,[254,1,“S”] ‘Stopportb.3 = 0pulsout portb.4, 668 ‘Right servo stoppulsout portb.5, 676 ‘Left servo stoppause 18portb.3 = 1goto mainleft:serout portb.1,6,[254,1,“L”, #recdata[2]]for x= 1 to 7pulsout portb.4, 696 ‘Right servo forwardpulsout portb.5, 676 ‘Left servo stoppause 20next xgoto main:right:serout portb.1,6,[254,1,“R”,#recdata[2]]for x= 1 to 7pulsout portb.4, 668 ‘Right servo stoppulsout portb.5, 648 ‘Left servo forwardpause 20next xgoto main:fwd:serout portb.1,6,[254,1,“F”,#recdata[8]]for x= 1 to 7pulsout portb.4, 696 ‘Right servo forwardpulsout portb.5, 648 ‘Left servo forwardpause 20next264 Chapter Fourteen goto main:bwd: serout portb.1,6,[254,1,“B”,#recdata[8]] for x= 1 to 7 pulsout portb.4, 640 ‘Right servo backward pulsout portb.5, 704 ‘Left servo backward pause 20 next x goto main: display:serin2 portb.0,84,20,main,[str recdata\3]for x = 0 to 3serout2 portb.1,16468,[“ ”,recdata[x]]next xpause 1500serout2 portb.1,16468,[254,1]returnRobot construction By the time this book goes to print, this artificial vision robot will be available as a kit from Images SI Inc. Visit the CMU camera website at http://www.cmucam.com. We begin by assembling two part A’s of the standard servomotor Figure 14.12 Two servomotor brackets, part A, assembled.Color Robotic Vision System 265 bracket together (see Fig. 14.12). A front U bracket is made to assemble to the front of the two part A’s (see Fig. 14.13). The inside width of the front U bracket is the same as the width of the CMU camera, approximately 2.125 in. The front U bracket has a hole near the front for the shaft of the front wheel.There are holes near the top front of the U brac ket also, not shown in the figure. Figure 14.13 Brackets with U bracket assembled. Figure 14.14 Robot base with servomotors , wheels, and multidirectional front wheel.266 Chapter Fourteen Figure 14.15 Finished robot. Next we assemble our two servomotors and front wheel onto the base assembly (see F ig. 14.14). The wheels for the servomotors are the same type of wheels used in Chap. 8. The front universal multidirectional wheel is the same one used in the Braitenberg vehic les in Chap. 9. Two small Lshaped mount ing ears are made to attac h the CMU camera to the front of the U brac ket. I constructed the entire circuit on a PIC Experimenter’ s Board. I changed the Xtal on the board from 4.0 MHz to 16 MHz. Power for the circuit ma y be obtained from an external power supply or an onboard battery power supply . The finished robot is shown in F ig. 14.15. Running the Program When you first run the robot,you may want to have it lifted so the wheels don’t touch. It’s a lot easier to c heck operation and function without ha ving to run after the robot. Use the experience you gained with object/targets using program 2. The LED D1 flashes after the auto light adjustment to signal you to put the target in front of the camera. The D1 LED also flashes when the robot is in the stop loop . I included the flashing LED because it’s not always easy to see the LCD displa y. The program reads the MMX value from the CMU camera and determines whether the robot should turn left or right. You can adjust these values to suit your particular target. Do not make the greater than (.) and less than (,) values of MMX too close. If you do, the robot will quiver left and right constantly. If you find the robot constantly overshooting when it turns to the left or right, you can reduce the loop value (x) in these turn subroutines .Color Robotic Vision System 267 Going Further Obviously, we have just scratched the surface of playing with the CMU camera. One feature I didn’t have time to implement was an upanddown tilt servomotor that uses the MMY parameter . This involves adding another servomotor to the robot, but would allow the robot to follow a target as it moves up and down. I quickly approached the memory limit of the PIC 16F84. If I had had more time, I would have implemented it, using another PIC microcontroller with a little more memory. The PIC 16F628 is port B–compatible with the 16F84 and has twice as much memory (2048 bytes). Latest updates and information on the CMU camera can be found at http://www.cmucam.com or http://www.cmucamera.com. Parts List CMU camera (2) Servomotors (HS425) (2) Part A servomotor brackets16MHz PIC Experimenter’s Board(2) Servomotor wheels PIC 16F84, 20 MHz16MHz crystal(2) 22pF capacitors (2) 330V, 1/W resistors 4 4.7kV, 1/W resistor 4 Multidirectional wheel Aluminum sheet metal, shaft, plastic screws, and nuts Available from Images SI Inc. (see Suppliers at end of book).This page intentionally left blank.SuppliersImages SI Inc. 109 Woods of Arden Road Staten Island, NY 10312 (718) 6988305 (718) 9826145 (fax) www.imagesco.com Jameco Electronics 1355 Shoreway Road Belmont, CA 94002 (800) 8314242 (800) 2376948 (fax) www.jameco.com JDR Microdevices 1850 South 10th Street San Jose, CA 95112 (800) 5385000 (800) 5385005 (fax) www.jdr.com 269This page intentionally left blank.Aluminum, 90Aluminum hip bar, 230Applics directory, 22AutoCodeCompletion, 23Behaviorbased (neural) architecture, 83–86Behaviorbased robotics, 86, 87Berger, Hans, 87Bigfoot walker, 225Binary, 65–68Binary counting program, 71–72Binary number table, 66Bipedal walker robot, 225–242assembly, 230–231feedback, 227finished product (photographs), 234, 235footpads, 228–230movie, 225parts list, 242pivoting, 242possible improvements, 241–242program, 233–241schematic, 231–233servomotor brackets, 228servomotors, 226, 227turning right/left, 242(See also Hexapod walkers)Bit, 65Bookmarks, 24Braitenberg, Valentine, 125Braitenberg vehicles, 125–141back wheels, 129–131CdS photoresistor cells, 133, 136–137first vehicle, 128–141front wheels, 131–132graphs (transfer functions), 127neural I/O relationships, 126parts list, 141IndexBraitenberg vehicles (Cont.):photograph of finished vehicle, 140second vehicle (avoidance behavior), 141sensor array, 137–13916F84 microcontroller, 139testing, 139, 141Breadboard, solderless, 49–50Brooks, Rodney, 86Building intelligence, 83–86button, 72–75Byte, 65Camera (see Color robotic vision system)cd, 16cd.., 17cd\, 17CdS photoresistors:Braitenberg vehicles, 133, 136–137 Walter’s turtle, 109–110Center punch, 89, 90CMU camera, 244–245CodeDesigner, 23–35connect EPIC programming board to computer, 33EPIC programming board software, 33–34features, 23–24free version, 24parts list, 35set the options, 25–33software installation, 25CodeDesigner Lite, 24Cold rolled steel (CRS), 90Colored PicBasic syntax, 24Color robotic vision system, 243–267CMU camera, 244–245finished product (photograph), 266improvements, 267incandescent/fluorescent lighting, 254–259Copyright © 2004 The McGrawHill Companies. Click here for terms of use . 271272 Index Color robotic vision system (Cont.): interfacing CMU camera to robot, 250–251parts list, 267program 1, 251–254program 2, 254program 3, 261–264robot construction, 264–266running the program, 266serial communication, 245–248servomotor for robot, 259–26116F84 runs at 16 MHz, 251VB application program, 248–250Compiler, 2installing PicBasic, 11–12installing PicBasic Pro, 12–18PicBasic (PicBasic Pro), 2–4using, 7Components (see Parts list)Connected speech recognition system, 168Consumables, 5 Continuous speech recognition systems, 168copy, 16Counting program, 71–72Creation of intelligence, 83–86CRS, 90dir, 16DOS commands, 16DOS programming, 37–48compile, 39–43 EPIC programming board software, 44–48 programming the PIC chip, 43–44 DOS prompt, 17DOS window, 17Drill, 90Edge detection, 243–2448bit number, 65Elmer and Elsie, 88EPIC Programmer, 3EPIC Programmer software/hardware, 3EPIC programming board software, 33–34,44–48EPIC programming carrier board, 4, 5Experimenter’s Board (see PIC Experimenter’sBoard) Expert system, 83–86 Finished products (see Photographs of finished products; Projects)Firmware, 5Flash memory, 5FlexiForce pressure sensor, 227Hammer, 90Hexapod walkers, 143–164center (tilt) servomotor, 154–155construction, 148–164electronics, 158–159leg positioning, 152–153linkage, 154, 155microcontroller program, 159–164mounting the servomotors, 151–152moving backward (backward gait), 146–147moving forward (forward gait), 145–146parts list, 164photograph of finished robot, 161robot base (diagram), 149robot legs (diagram), 150sensors, 155–158servomotors, 144, 145tripod gait, 143–144turning left, 147–148turning right, 148(See also Bipedal walker robot)Highlevel language, 2HiTec HS322HD servomotors, 227HiTec HS425BB servomotor, 90–97HiTec servomotor horn, 189HM2007 speech recognition integrated circuit, 165HS85MG servomotors, 215HS322HD servomotors, 227HS425BB servomotor, 90–97Image processing, 243(See also Color robotic vision system)Images SI Inc., 269“Imitation of Life, An” (Walter), 88Input, 72, 77–78input, 77–78Installation:CodeDesigner software, 25EPIC software, 19–22firmware, 7PicBasic compiler software, 11–12PicBasic Pro compiler, 12–18Intelligence, 83–86Isolated speech recognition systems, 168Jameco Electronics, 269JDR Microdevices, 269Label listbox, 24Layered behavioral responses, 85–86LCD display, 54–56, 60–65Legged walkers, 143(See also Bipedal walker robot; Hexapodwalkers)Index 273 Line error highlighting, 23Living Brain, The (Walter), 88Machina Speculatrix, 88“Machine That Learns, A” (Walter), 88md, 16Metalworking tools, 89–90Microchip Technology Inc., 1, 2Microcontroller:defined, 1features, 116F84, 5–6why used, 1–2microEngineering Labs, Inc., 2, 4Models (see Projects)Neural (behaviorbased) architecture, 83–86Nibbler, 89, 90Output, 70–71, 77–78 output, 77–78 Parts list, 8–9bipedal walker robot, 242Braitenberg vehicle, 141CodeDesigner, 35color robotic vision system, 267hexapod walker, 164robotic arm, 223–224speech recognition system, 182–184testing circuit, 79, 81Walter’s turtle, 123–124path, 16Pavlov, Ivan, 87Pavlov’s dogs stimulusresponse experiment, 87peek, 75–77Photographs of finished products:Braitenberg vehicles, 140color robotic vision system, 266hexapod walker, 161robotic arm, 222, 223speech recognition system, 179–181Walter’s robot, 122PicBasic, 2PicBasic compiler, 2–4PicBasic Pro compiler, 2–4PicBasic Pro/PicBasic softwarepackages/manuals, 3PIC chips (PICmicro chips), 2PIC Experimenter’s Board, 56–65builtin LCD, 60–65experiment (prototyping area), 58–60LCD display, 54–56, 60–65test circuit (Walter’s turtle), 111PIC Experimenter’s Board (Cont.): use, 56–58PICmicro chips, 2PIC programming overview, 2–4PIC 16F84 circuit, 109PIC 16F84 microcontroller, 5–6poke, 70, 71Port A register, 69Port B register, 67Port registers, 65–70Pressure sensor, 227Programming carrier board, 4, 5Programming the PIC chip, 7Projects:bipedal walker robot, 225–242Braitenberg vehicles, 125–141color robotic vision system, 243–267finished products (see Photographs of finished products)hexapod walker, 143–164robotic arm, 185–224speech recognition systems, 165–183Walter’s turtle, 87–124(See also individual subject headings)QuickSyntaxHelp, 23Redo/undo, 24Registers, 65poke, 70, 71port, 65–70port A, 69port B, 67TRIS, 68 TRISA, 68, 69TRISB, 67, 68Robotic arm, 185–224 adding a base, 216–220 assembly multipleservomotor assemblies, 189–192finished product (photograph), 222, 223parts list, 223–224servomotor brackets, 185–189servomotor controllers, 199–215servomotors, 197–199, 215Robotic arm gripper, 193Robotic arm gripper assembly, 195Servomotor:bipedal walker robot, 226, 227centering program/circuit, 78–80color robotic vision system, 259–261hexapod walker, 144, 145photograph of, 79 274 Index Servomotor (Cont.): Testing the PIC microcontroller (Cont.): robotic arm, 197–199, 215 PIC Experimenter’s Board, 56–65 Walter’s turtle, 90–97 (See also PIC Experimenter’s Board) what it is, 78 registers (see Registers) Servomotor bracket kit, 186 schematics of test circuit, 50–53 Servomotor bracket travel, 187 servomotors, 78–79 Servomotor horn, 189 (See also Servomotor) Shears, 89, 90 solderless breadboard, 49–50 Sheet aluminum, 90 troubleshooting the circuit, 54 Sixlegged walker robot, 143–164 Tilden, Mark, 86 (See also Hexapod walkers) Tripod gait, 143–144 16F84 PIC microcontroller, 5–6 TRISA register, 68, 69 Solderless breadboard, 49–50 TRISB register, 67, 68 Speakerdependent speech recognition system, TRIS register, 68 167–168 trisx.x, 70 Speakerindependent speech recognition system, 168, Turtle robot (see Walter’s turtle) 172–173 Speech recognition systems, 165–183 Undo/redo, 24 applications, 167 circuit construction, 168–169 Vehicles—Experiments in Synthetic Psychology clearing trained word memory, 172 (Braitenberg), 125 components, 165–166 Vise, 90 finished product, 179–181 listening/speech understanding, 167 Walker robots (see Bipedal walker robot; Hexapod microphone, 181 walkers) parts list, 182–184 Walter,William Grey, 87–88 programming speech recognition circuit, 177–178 Walter’s turtle, 87–124 16F84 microcontroller program, 176–177 attaching bumper to robot base, 101–103 software approach, 167 behavior, 120–121 speakerdependent/speakerindependent system, bumper switch, 104–106 167–168, 172–173 center of gravity, 100–101 speech interface control circuit, 173–176 drive/steering motors, 90 styles of speech, 168 equipment, 89–90 testing recognition, 172 fudge factor, 121 training the circuit, 169–172 handedness, 123 voice security system, 173 light intensity, 121–123 wireless robot control, 181 modes of operation, 88 Speech software, 167 mounting steering servomotor, 107–108 Stainless steel, 90 observed behavior, 88–89 Standalone servomotor bracket assembly, 191 parts list, 123–124 Standardbased expert programming, 83–86 photographs of finished product, 122 Statement description, 23 photoresistor, 109–111 Statement Help, 23 power, 119–120 Subdirectory, applications, 22 program, 115–119 Suppliers, 269 schematic, 114–118 sensor array, 112–114 Testing the PIC microcontroller, 49–81 servomotor, 90–97 accessing ports for output, 70–71 sheet metal fabrication, 97–98 button, 72–75 shell, 99–100 counting program, 71–72 sleep mode, 119 input, 72 Windows version of EPIC software, 8 input/output commands, 77–78 wink.bas circuit (solderless breadboard), 53 LCD display, 54–56, 60–65 wink.bas program, 45 parts list, 79, 81 peek, 75–77 xcopy, 16ABOUT THE AUTHOR John Iovine is the author of several popular TAB titles that explore the frontiers of scientific researc h. He has written Homemade Holograms: The Complete Guide to Inexpensive, DoItYourself Holography; Robots, Androids, and Animatrons: 12 Incredible Projects You Can Build, considered a cult classic; Kirlian Photography: A HandsOn Guide; Fantastic Electronics: Build Your Own NegativeIon Generator and Other Projects; and A Step into Virtual Reality. Mr. Iovine has also written extensively for Popular Electronics, Nuts & Volts, Electronics Now, and other periodicals.This page intentionally left blank.