Application of Sensors for Measurement of Tissue PropertiesIntroductionSensors in biomedical applications enable detection of biological events and convert them into signals, thus serving as an interface between a biological and information system (Harsányi, 2001). Applications in minimally invasive surgery (MIS) require a high degree of accuracy, which could be achieved by a robot or manipulator performing selected tasks during surgery. Common sensors such as CT, MRI, US, or x-ray cannot acquire extensive data that is accurate and up-to-date for use with a robot (Stallkamp & Schraft, 2005). Applications have focussed on properties of biological tissues, as they contain useful information such as age, gender, health of organs, etc. Young’s modulus or stiffness, Poisson’s ration, and viscosity are some properties. Young’s modulus is the most important as it depends on the composition of tissue. MIS involves performance of surgical procedures on internal organs using small incisions. Surgical mechanisms and endoscope are inserted for the manipulation of tissues and viewing the surgical site. Images of internal organs and surgical instruments that have been inserted are viewed using a monitor. Reduction of trauma, reduced risk of inflammation, reduction in post operative complications, and fast recovery are advantages of MIS. Reduced dexterity, loss of tactile sensitivity, and force feedback are disadvantages of MIS. For effective performance of MIS, the surgeon should be able perform controlled manipulation of tasks such grasping of internal organs, transfer of load during lifting, removal and suturing of tissues, etc. This requires the ability to feel the tissue, and sense the blood vessels and ducts during surgery (Dargahi et al., 2000).Advances in Sensor TechnologyGray and Fearing (1996) studied a surface micromachined microtactile capacitative sensor array, and concluded that the sensors possessed good spatial uniformity, were capable of detecting millinewton forces, and there was good interpolation between elements. Dargahi et al. (2000) demonstrated that the magnitude and applied force could be determined from magnitude and slope of signals from Polyvinlyidene fluoride (PVDF) sensing elements. The prototype sensor exhibited high sensitivity, large dynamic range, and high signal-to-noise ratio. The sensor could be integrated with an endoscopic grasper. Hwang et al. (2003) studied surface-micromachined flexible polysilicon sensor array; and concluded that the sensor had greater sensitivity than metal strain gauge. There were hysteresis problems, but there were no detectable proximity effects. Dargahi and Najarian (2003) studied semiconductor microstrain-gauge sensors, and concluded that the sensor had the capability to measure with reasonable accuracy the magnitude and position of an applied load. The electronic device could be integrated into and endoscopic grasper. Tavakoli et al. (2005) studied haptic interaction in robot-assisted endoscopic surgery; and concluded that haptic feedback makes robotic surgery more efficient, accurate and reliable, and end effectors were capable of using detachable and disposable tips. Mirbagheri and Dargahi (2005) conducted finite element analysis of a membrane-type piezoelectric tactile sensor with four sensing elements; and concluded that the tactile sensor could measure the magnitude of applied force and estimate shape of objects contacted, and the sensor could be used in MIS in endoscopic tools and tactile imagers. Najarian et al. (2006) proposed the design and fabrication of piezoelectric-based tactile sensor, which could be applied to measure total force on a sensed object and compliance of the tissue or sensed object. Hosseini et al. (2006) presented a computational tactile sensing approach for the detection of tumours, and demonstrated that characteristics of tumour could be predicted by the application of tactile sensing. Najarian et al. (2006) demonstrated that stiffness of sensed objects could be measured with reasonable accuracy (20 percent error) for forces within 0.1-1 N, and showed that the endoscopic tactile sensor prototype could be used in robotic procedures for performing minimally invasive surgeries. Nieminen et al. (2006) studied acoustic properties of articular cartilage under mechanical stress; and concluded that volumetric changes during compression were relatively small in human cartilage, and mechano-accoustic techniques were not significantly impaired. Jalkanen et al. (2006) studied prostate tissue stiffness as measured with a resonance system. The variation in tissue stiffness caused by change in resonance was detected by the system. Frequency change and force were parameters used to detect stiffness in silicone rubber and human prostrate tissue in vitro. Silicone rubber variations could be detected. Prostrate tissue measurements showed reproducible difference between tumour and healthy tissue. Variations between tumour and normal tissue were assumed to be caused by differences in tissue composition. Liu and Ebbini (2007) studied viscoelastic property measurement in thin tissue constructs using ultrasound; and concluded that acoustic radiation force (ARF) methods showed significant improvement while compared to conventional ultrasound. Presence of nearby boundary conditions such as skin imaging makes it difficult to interpret ARF induced displacements. A dual element transducer was developed for generating and tracking localized tissue displacements. Elastography phantoms on a hard substrate that mimicked tissue produced spatio-temporal maps. Resonant behaviour was exhibited by the frequency response determined by stiffness and thickness. Extended Kaufman filter was used for tracking apparent stiffness and viscosity of samples. Rago et al. (2007) studied new developments in ultrasound for predicting malignancy in thyroid nodules; and concluded that US elastography could be used to distinguish malignant lesions from benign lesions, and could be used as an adjunct tool for the diagnosis of thyroid cancer. Jalkanen et al. (2007) studied spatial variations in prostate tissue histology as measured by a tactile resonance sensor, and concluded that tumours around the tip of the sensor could be detected. A clinical device based on resonance sensor could be developed for detecting tumours and guiding biopsies. Liu et al. (2007) proposed rolling mechanical imaging: a novel approach for soft tissue modelling and identification during minimally invasive surgery. The device could be used in robotic assisted surgery. The mechanical rolling image could be used to obtain prior information for selection of insertion point for needle and improving accuracy of trajectory prediction. Zbyszewski et al. (2008) proposed wheel/tissue force interaction: a new concept for soft tissue diagnosis during MIS. The tactile air-cushion sensor sensor enables acquisition of tactile information over soft tissue rapidly over large areas. Zbyszewski et al. (2008) developed a force sensor for localizing tissue abnormalities, and provide tactile feedback during MIS. An air supported ball has been attached at the end of a tubular shaft for indenting tissue under investigation. Changes in tissue stiffness cause changes in position of the ball, which is measured by an optical sensing system. Tactile information could be rapidly acquired over a large area of soft tissue. The performance was similar to the wheel-based force sensor.Sensor TechnologyThe type of application and the type of object to be contacted have motivated the design of a variety of tactile sensors. Complexities arise when tactile sensors are directed towards biological tissues, requiring more sophisticated designs. Tactile sensors could be classified as mechanical, capacitative, magnetic, optical, piezoelectric (PZT), piezoresistive, and silicon-based. Special silicon-based tactile sensors have been developed, which integrate signal processing with sensors. Diagnostic and therapeutic applications of robotic surgery include robotic minimally invasive surgery in areas including heart surgery, microsurgery, medical telemetry, endoscopy, urology, neurosurgery, arthroscopy, and orthopaedic surgery (Dargahi & Najarian, 2005). Robotic surgery has presented many diagnostic and therapeutic possibilities including MIS. Tactile or visual sensing is important for MIS. Displacement-based Sensor The change in the length of a spring is indicative of the magnitude of force acting on it according to the force-displacement relationship. Potentiometer, digital encoder, and a linear variable differential transformer are displacement sensors that can accurately indicate displacement. Force value could be estimated by the application of a calibration rule. Low forces as low as millinewton could be accurately measured by LVDT based transducers. Advanced techniques include the simulation of characteristics in a servo stystem. Friction at joints and backlash in drive mechanisms could cause errors in force estimation. Also, mechanical properties such as tissue hardness could be felt (Puangmali et al., 2007). Current-based SensorCurrent of motors actuating joints could be used to measure forces. Joints driven by servo motor have torques or forces proportional to armature current of the motors. The force on the axis of a motor is a disturbance of servo control system. Increasing the motor current for producing force to balance the disturbance effects is interpreted as the force under measurement. A calibration rule is applied to calculate the force. Since the force is measured indirectly, the measure is not very accurate (Puangmali et al., 2007). Pressure-based SensorPneumatic actuator is a method to develop a force sensing device. Pneumatic driven forceps have joints driven by wire power transmission mechanism, which are actuated by servo pneumatic actuators. Introduction of a disturbance observer to the system constitutes the force sensing capability. A neural network has been used to estimate driving forces for all joints. Effective estimate of the forces requires comparison with sensed forces from pressures of pneumatic cylinders, and low pass filtering. The quality of estimation could be degraded by distance of measurement from point of application, backlash of driving mechanism, friction variation at joints and cylinders. Contact forces with resolution of 0.1 N could be determined (Puangmali et al., 2007). Resistive-based SensorStrain gauges have been used as sensing elements in force measuring devices. Flexure is a flexible structure used for boding with the strain gauge for providing an accurate means of force measurement. Resistance of a strain gauges changes when a strain is caused by a force, which could be measured. However, there lies a trade-off between the stiffness and sensitivity. Higher stiffness provides low measurement sensitivity, causing difficulties in force measurement. Hysteresis is another phenomenon that restricts the flexibility of the sensing device. Forces with resolution of 0.25N in axial 0.05N in radial direction have been measured in applications. Force and torque measurements as high as 20 N and 200 N-mm respectively have been developed (Puangmali et al., 2007). Capacitive-based SensorA powerful means for the detection of extremely small deflections in structures is capacitive sensing. Capacitive tactile sensors have flat and parallel plates that are separated by high dielectric material (Gray & Fearing, 1996). Capacitive sensing has good sensitivity to small deflections without temperature dependence. Miniature tactile array sensors for endoscopic tools have been developed. A silicone rubber layer covers the array for spatial distribution of forces, and low passes mechanical filtering. The array is less than 1mm2. Contact force cause deformations, which cause change in capacitance of sensing elements, and are detected by the use of oscillatory circuits. The ability to resolve millinewton forces, over a range of 2N forces and its small size has made it attractive for MIS devices. Variations in pressure of arteries, stiffness, and texture of soft tissue could be measured. The spatial distribution of forces over sensing surfaces could be detected by tactile sensors, allowing measurement of variations in shape and hardness. Precise measurement over a small range is the most suitable application for capacitive based tactile sensors (Puangmali et al., 2007). Acoustic-based SensorWell fabricated piezoelectric materials have been considered effective sensing elements for tactile sensors. Large output voltage could be generated for a small deformation, which distinguishes them from most other sensors. The sensing elements do not require electricity, their operation is reliable, and have a wider range of applications. PVDF has been used for the construction of tactile sheet sensors. When a PVDF film is stressed by the application of a force, voltage signals are generated from polarization charges. Magnitude, position, and distribution of applied force could be determined by the measurement of the amplitude of output voltages at electodes, and comparison. The sensor has exhibited good linearity and high sensitivity. Other materials include ferroperm piezoceramics. Applications have displayed high signal-to-noise ratio for 2N range, and applications have been developed for measurement of tissue forces upto 15N. Charge leakages and changes in temperature are limitations of this type of sensors (Puangmali et al., 2007). Optical-based SensorOptical sensors use fibre optic cables for carrying force information from sensing region to opto-electronic equipment. A light source, transduction element, and optical detector are components of an optical sensor. Light in proportion to force measured is modulated by the transduction element, which travels through the fibre to the detector. The light is converted to an electric signal for further processing. Applications have been developed with resolution of 0.04N, and ranges of 2.5N in axial direction and 1.7N in radial direction. The sensor is immune to electromagnetic interference, and has compatibility with magnetic resonance. Limitations are degradation of accuracy of measurement by lack of compensation, and some optical fibres are unsuitable as wires (Puangmali et al., 2007). A summary of sensor characteristics and suitability of applications has been illustrated in table 1.Table 1. Sensor CharacteristicsSensor TechnologyApplicationCharacteristics ConclusionDisplacement-based sensorPotentiometer, digital encoder, LVDTResolution: milli NFriction at joints and backlash in drive mechanisms could cause errors in force estimation.Current-based sensorMotors actuating jointsSince the force is measured indirectly, the measure is not very accurate.Pressure-based sensorPneumatic actuatorsResolution: 0.1NThe quality of estimation could be degraded by distance of measurement from point of application, backlash of driving mechanism, friction variation at joints and cylinders.Resistive-based sensorStrain gauge sensorsResolution: 0.25N axial, 0.05 N radial, Range: 20NHigher stiffness provides low measurement sensitivity, causing difficulties in force measurement. Hysteresis is another phenomenon that restricts the flexibility of the sensing device.Capacitive-based sensorCapacitive sensorsResolution: 0.001 N, Range: 2NThe spatial distribution of forces over sensing surfaces could be detected by tactile sensors, allowing measurement of variations in shape and hardness. Precise measurement over a small range is the most suitable application for capacitive based tactile sensors.Acoustic-based sensorPVDF sensorsRange: 15NCharge leakages and changes in temperature are limitations of acoustic of sensors.Optical-based sensorOptical fibre sensorsResolution: 0.04 N, Range: 2.5N axial, 1.7N radialThe sensor is immune to electromagnetic interference, and has compatibility with magnetic resonance. Limitations are degradation of accuracy of measurement by lack of compensation, and some optical fibres are unsuitable as wiresSelected Sensors for Measurement of Tissue PropertiesGrasping and manipulation of tissues during robotic surgery requires information such as the magnitude and position of force applied to objects. Requirements of tactile sensors include small size (less than 1mm square), inexpensive and disposable, mass fabricated, and packaged usefully (Gray & Fearing, 1996). Tactile sensors should have the capability to determine the pressure distribution; the magnitude and position of applied force between the sensor and tissue (Dargahi et al., 2000). Sensitivity of tactile sensors strongly depends on several variables that are determined by the physical and chemical properties of the sensor. Force sensitivity within the range of 0.1 to 11 N has been considered satisfactory for medical applications. A fundamental requirement for safe handling of biological tissues is the measurement of magnitude and location of forces applied by the endoscopic grasper (Dargahi & Najarian, 2003).Silicon has piezoresistive properties that have been used in sensor and actuator applications. Piezoresistive properties include changes in electrical resistivity under deformation, allowing transducers to be built for mechanical and chemical measurements. The ratio of change of resistivity R induced by change in length of resistor l has been defined as the piezoresistive coefficient π, and is given by . π = 2 for many metals caused by geometrical deformation, and ranges from -120 to + 120 caused by doping, temperature and orientation of crystal. It is believed that thin silicon and patterned polysilicon film could be used as strain gauges (Hwang et al, 2003). Compressive stiffness of elastic material is characterized by Young’s modulus, which could be determined acoustically by the measurement of speed of sound and material density. Hooke’s law is given by: , where is the stress tensor, is the elastic stiffness matrix, and is the strain tensor. Young’s modulus (E) and Poisson’s ratio (v) are independent elastic constants for a perfectly isotropic material. Young’s modulus, aggregate modulus (HA) and shear modulus (µ) are given by: The speed of sound wave (c) in isotropic elastic solid is given by: where v and are Poisson’s ratio and density of material. Modulus and speed in a narrow bar is given by: E could be determined by the measurement of c when is known. US reflection from an acoustic interface depends on acoustic impedance and is given by: where R is the reflection coefficient, , i=1,2. This, E could be determined by the measurement of R when are known (Saarakkala et al., 2004).Piezoelectric tactile sensors have been widely applied for robotic surgery. Piezoelectric sensors are based on piezoelectric material properties, and their capacitance. A piezoelectric material subjected to stress results in the accumulation of electric charge on the sensing elements or capacitors, which could be measured and processed. Piezoelectric sensors have been designed with a large number of sensing elements in matrix arrangement. However, this results in crosstalk or unwanted response from adjacent elements, causing measurement errors, requirement of large number of cables, and fragility. An alternative approach to the design of sensors is the use of thin membrane with few sensing elements. Sensor configuration, basic equation and material properties, closed-form solution have been analyzed by Mirbagheri and Dargahi. The application of ultra thin PVDF that has a radius of 45 mm and 25 µm thickness has been analyzed. Four rectangular 2 mm wide aluminium electrodes were located at a radius of 10 mm from the centre two on drawn direction 1. They could be formed by vacuum deposition on PVDF film or etching aluminium layer deposited on PVDF. The intersection of each set of two electrodes on either side of the membrane resulted in a square shaped sensing element. It was unsupported to get a better perception of PVDF film (see table 2) behaviour. The circular film circumference is fully constrained (Mirbagheri & Dargahi, 2005). Table 2. PVDF Film Material Properties Manufactured by Goodfellow Company, USASl #AttributePropertyYoung’s modulus (drawn and transverse directions)2 GPaPoisson’s ratio0.34Density1760 kg/m3Tensile strength (drawn and transverse directions)180 MPaPiezoelectric constant (drawn and traverse directions)8 pC/NYield strength (both directions)45There are piezoelectric coefficients along drawn, transverse, and thickness directions, denoted as d31, d32, and d33. Charge contribution caused by deformation in thickness direction 3 has been ignored. The resultant output charge on each sensing element is given by the expression , where d31 and d 32 are piezoelectric coefficients, σx and σy are stresses acting along the Cartesian coordinates x and y, and Sel is the electrode area. The average stresses on the sensing elements are converted to electrical charges and voltages using the formula described earlier. Stress could be obtained using the closed-form solution or finite element analysis. Closed form solution requires the determination of out-of-plane deflections and strains, derivation of stresses developed in the membrane by membrane theory and mechanics, and transform them into electrical charges based on piezoelectric properties. Numerous assumptions are made to limit the complications presented by extraction of closed form solution. Computations required include strain, stress and finally output charge. Asymmetric loading and PVDF being an anisotropic material, obtaining a closed-form solution is cumbersome. In the finite element method, sensor behaviour with reference to magnitude and position of applied forces could be mapped by drawing arcs with radial increments (for example 4 mm) and angular lines with subdivisions (for example 10 degree starting at 45ͦ to 85ͦ). Load is applied at each intersection point as pressure acting on circle radius (for example 1mm). The contact of a load with a sensor acts like an impulse as the application is sudden and lasts for a short time. Mirbagheri and Dargahi selected harmonic, sinusoidal load with 0.25N magnitude and frequency of 15 Hz for ease of application in their proposed design. The response of a harmonic load is easier to record than an impulse load. Static and dynamic loads exhibit a similar behaviour. The area could be meshed with a sensing element (for example Shell63), and the choice should be based on performance in simulation of membrane behaviour. Mapping could be performed with reference to element A, and application of load at prescribed points. The charges on all elements are calculated and plotted. This gives an idea of system response to various loading conditions. A linear structure’s steady state response to varying sinusoidal loads could be determined by harmonic response analysis. The sensing elements are filled with elements and nodes on meshing. Computation includes the average of stress components developed at the nodes of each sensing element, which are substituted in the equation described earlier for determination of charges. There is a decline in charge as load application moves away from the element. The closer a load is to an element, there is an intensification of the stress concentration surrounding it, resulting in a higher amount of charge. The triangulation method is used to relate the position to the magnitude of the load. This result in an isocharge contour; locus of points, which when subjected to load results in the accumulation of charge on its sensing element. The information is stored in a database, and when each time a load is applied, the database information is matched with resulting contours. Shape estimation is another application of the sensor. Robotic surgery requires knowledge of the outer contour of the object and its centroid for gripping and manipulating objects. The sensor allows the determination of irregularities such as sharp corners and two dimensional objects. Stress concentration increases in the area engulfing the element, giving a sense of the outer boundary. The location of the centroid of the shape is guessed by comparing the charges on elements (Mirbagheri & Dargahi, 2005). Tactile properties are of critical importance in MIS. However, loss of tactile capabilities is a major drawback in MIS. Tactile sensors are lacking in commercially available endoscopic graspers. Strain gauges have been applied for the measurement of the magnitude of applied force. Objects with different elastic properties could be identified with the sensor. Najarian et al. (2006) have designed an endoscopic piezoelectric tactile sensor for the measurement of applied force, and compliance of tissue or sensed object. A piezoelectric sensor consists of a transduction point made of micro-electric puller mounted on a micro-manipulator platform. Relative deformation the tissue or object contacted on compliant rigid objects is the principle behind detection of sensed object compliance. This includes the placement of PVDF film between a rigid cylinder and a plate that has been used for measurement of force applied on the rigid element. A second PVDF film is included between two base plates that are used for the measurement of total force applied on the sensor. The PVDF films response for different load sets enables the measurement of compliance of the sensed object. The purpose of teeth while holding an object is served by the rigid cylinder. Good holding properties could be achieved by increasing the number of rigid cylinders. Integration of this sensor with an endoscopic grasper requires miniaturization (Najarian et al., 2006). A tactile sensor made of PVDF membrane was applied in a method for the determination of existence of embedded object within biological tissue. Tactile images and stress graphs were extracted, and tissue modelled (Hosseini et al., 2006).Transfer of momentum from an acoustic wave to a propagation medium causes ARF. The ARF produced in a homogenous attenuating medium that is unbounded is given by where c is the speed of sound in the medium, α is the phantom absorption coefficient, and I is the acoustic beam temporal intensity. The net force produced in the presence of boundary conditions and radiation pressure at the boundary is given by , where subscripts refer to sides of the reflecting boundary, A is the cross sectional area, and R the reflection coefficient. The standard forced harmonic oscillator is used for estimation of viscoelastic tissue properties, which is given by + 2ε , where is the undamped natural frequency, and /2Mω0 is the damping coefficient, K is the stiffness, and is the viscosity of the tissue. is the process noise that accounts for the uncertainty in applied ARF and other contributing forces. Process noise is accounted for by the Kalman filter. Accelaration, velocity, ω0 , and ε are state variables. Piezoelectric resonance is used to sense stiffness of soft object. Lead zirconate titanate (PZT) is a widely used piezolelectric element in the sensor, comprising of transducer part and pick-up part. An alternating voltage close to the resonance frequency of the PZT element is applied on the transducer to vibrate the element. The pick-up detects the vibration, which is processed by a feedback circuit ensuring that the driving frequency is same as the resonance frequency. Δf is the change in resonance frequency given by a zero phase shift sensor system condition, for oscillation when the sensor is in contact with the object. Change in resonance frequency is given by acoustic impedance of the object given by , where β is the reactance part of acoustic impedance, v0 is sound velocity in PZT element, Z0 is acoustic impedance, and l is length of the element. is dependant on contact area between sensor tip and measured object (Jalkanen et al., 2006).US indentation techniques provide information on tissue thickness, and enables calculation of material properties. A prototype of arthroscopic US indentation instrument was developed capable of measuring stress, strain and tissue thickness for the determination of articular cartilage dynamic modulus. In situ calibration enabled simultaneous determination of tissue equilibrium Young’s modulus and US speed. US attenuation in time domain was given by: , where is cartilage thickness, and A1(t) and A2(t) are peak to peak amplitudes of US pulse. Attenuation in frequency domain is given by: , where and are amplitude spectra of pulses, is the frequency range. US speed is given by: , where TOF(t) is the time of flight of US pulse. US speed using in situ calibration is given by: , where TOF(0) is the time of flight at the beginning of measurement. Constant acoustic properties for cartilage are not significantly impaired (Nieminen et al., 2006).Resonance sensors are based on PZT resonance. They have been used in stiffness measurements for detection of pathological conditions. When a sensor contacts a tissue with constant force, there is a change in frequency of the resonating system based on object stiffness. The sensor detects change in tissue stiffness (Jalkanen et al., 2007). Tactile sensors have been used to sense a range of stimuli, such as presence or absence of tissue, mapping of tactile image, or artificial palpitation. Information about the state of gripping or manipulation of tissue depends on factors such as force or position. Detection of tactile properties such as stiffness, temperature, and surface texture are key roles in tactile sensing for MIS (Najarian et al., 2006). Elastography is a dynamic technique for the estimation of tissue stiffness by the measurement of degree of distortion, when an external force is applied. During US elastography, two ultrasonic images are obtained and tissue displacement is tracked by assessment of propagation of the image beam (Rago et al., 2007).Intuitive Surgical has developed daVinci Surgical System with improvements such as stereoscopic view, and enhanced distal dexterity. Techniques for measuring tissue properties by force measurement include 1-DOF probing, stretching or grasping a tissue sample, and tactile sensor pressed on the surface of tissue. The effectiveness of these techniques were limited due to variable properties of soft tissue. The presence of an underlying abnormality could be determined by the examination of a large area of an organ. Size restrictions associated with MIS made it impractical to probe a large area in short time. Lui et al. (2008) developed a force sensitive wheeled probe for the identification of tissue parameters and abnormalities during MIS. The probe allows the surgeon the ability to probe solid organs to classify tissue properties and identify abnormalities in short time. The sensor is rolled over multiple paths at specific indentation depths for stitching individual force profiles and generating a mechanical image to indicate tissue stiffness. The indentation depth in increased over successive passes for assembling images. The data is analyzed for extracting Force-Tissue deflection characteristics of the tissue sample. Temporal and spatial variation of force signal could be measured by this technique overcoming limitations of current techniques.A wheeled probe was attached to the distil tip of a Mitsubishi RV-6SL 6-DOF robotic manipulator, and an ATI NANO17 Force/Torque sensor was attached between the sensor and the tip for measure force components; Fx, Fy and Fz imparted on the wheel by the tissue. An aluminium wheel grooved with 12 teeth was used as end-effector. A shaft mounted in ball bearing 2 connected the wheel to the sensor allowing the probe to follow trajectories along the x-y plane. The wheel centre and force axis was offset by 2 mm for provision of sufficient torque to rotate the probe while adjusting trajectory. A silicone phantom was used to gauge the efficacy of the wheeled probe while detecting underlying areas of tissue structure. Silicone phantom with six simulated rubber nodules was constructed (Liu et al., 2008).Tissue viscoelasticity parameters could be identified by 1-DOF. The ratio between the 1-DOF indentation force Fin and rolling normal force Fn is the linear function of rolling-normal force given by the expression: , where P1 = -0.04, P2 = 0.58, , Fro is the resultant force, ϵ represents the N/R ratio (rolling force over resultant force). Force-deflection characteristic on a sample point is required for prediction of behaviour of tissue sample in response to insertion of needle. Force-deflection curve profile is described as , where F is the force, u is indentation depth, and C,D,E are real numbers that are non negative, , where fui is the rolling force on rolling image with indentation depth (Liu et al., 2008). Mechanical Imaging involves the visualization of soft tissue by detecting mechanical stresses on an organ surface by the use of tactile sensor array. Mechanical imaging devices use tactile sensor array and a system for positioning. In applications for the diagnosis of breast and prostate cancer, palpitation has been an effective method in the detection and monitoring of pathological changes. Tactile imaging for comparison of nodule size estimate from preoperative examination with postoperative measurement has shown good repeatability and lesser error. However, adaptation of tactile sensors to MIS has miniaturization and sterilization problems (Liu et al., 2008). ConclusionA brief summary of studies has been illustrated in table 3.Table 3. Summary of StudiesReferenceStudySensor CharacteristicsConclusionDargahi & Najarian (2003)An Endoscopic Force-position Sensor Grasper with Minimum SensorsPrototype endoscopic grasper. Semiconductor microstrain gauges (10 ESB-020-120, Intertechnology). Maximum load: 10N. 0.5N increments.The sensor exhibited high sensitivity to force, good linearity, and large dynamic range.Dargahi & Najarian (2005)Advances in Tactile Sensors Design/Manufacturing and Its Impact on Robotics Applications – A ReviewManufacturing techniques of tactile sensors. PVDF based sensors, catheter micro tactile sensor (PZT based), micro needle sensorRobots need to grasp objects such as tissues and move them. Tactile sensors play an important role in improving robot performance. Dargahi et al. (2000)A Micromachined Piezoelectric Tactile Sensor for an Endoscopic Grasper—Theory, Fabrication and ExperimentsPrototype sensor with 3 layers; micromachined silicon, 25 µm thick PVDF film, and plexiglass. Vibration unit for applying forc: 15 Hz sinusoidal signal. Maximum load: 2N.Magnitude and position of force applied could be determined from PVDF signals of sensing elements.Gray & Fearing (1996)A Surface Micromachined Microtactile Sensor ArrayDesign of a capacitive sensor, doped polysilicon, air gap dielectric, 0.5µm, E: 160GPa, Resolution: milli N The sensor exhibited good spatial uniformity, and good interpolation between elements. There were hysteresis problems, but no proximity effects.Hakulinen et al. (2006)Ultrasonic Characterization of Human Trabecular Bone MicrostructureUS system (UltraPAC,Physical Acoustic Co.), 500MHzA/Dboard, 0.2–100 MHz pulser–receiver board. Ultrasound measurements, five pairs of focused transducers,range of frequencies (0.2–6.7 MHz). Pulse lengths 3.44 μs, 1.60 μs, 0.59 μs, 0.39 μs and 0.42 μs, centre frequencies of 0.5 MHz,1 MHz, 2.25 MHz, 3.5 MHz and 5 MHz. Quantitative US techniques and microstructural parameters had significant relationships that depended on frequency. US speed and structure had strongest association at 5MHz, in comparison to attenuation, reflection and backscatter. Harsányi (2001)Sensors in Biomedical ApplicationsA broad summary of fields of sensor application, and their technologies.Sensors serve as an interface between a biological and information system.Hosseini et al. (2006)Detection of Tumours Using a Computational Tactile Sensing ApproachMethod for modelling tissue with simulated tumour, and FEM for assessment of shape, size, depth, location3D analysis is a novel method for estimating characteristics of tumour, which could be applied to tactile sensing.Hwang et al. (2003)Surface-Micromachined Flexible Polysilicon Sensor ArrayFlexible polysilicon strain gauge, 4-inch silicon wafer, 2 µm phosphorous silica film, 2µm silicon nitride film.The sensor is more sensitive than metal strain gauge.Jalkanen et al. (2006)Prostate Tissue Stiffness as Measured With a Resonance System: A Study on Silicone and Human Prostate Tissue in VitroResonance sensor system (Venustrom, Axiom Co.), 200 Hz, 2-mm impression depth, speed 1 mm/sStiffness variation in silicone and human prostrate tissue in vitro could be detected by the system.Jalkanen et al. (2007)Spatial Variations in Prostate Tissue Histology as Measured By a Tactile Resonance Sensor.Resonance sensor system (Venustrom, Axiom Co.), 200 Hz, 2-mm impression depth, speed 1 mm/sSpatial variations in prostrate tissue histology were detected by the sensor, indicating tumours around sensor tip could be detected.Kiviranta et al. (2008)Indentation Diagnostics of Cartilage DegenerationIntegrated force device, flat miniature transducer (ø = 3 mm)Arthroscopic indentation instruments could be used for quantitative evaluation of mechanical properties of cartilage.Laasanen et al. (2002)Novel Mechano-Acoustic Technique and Instrument for Diagnosis of Cartilage DegenerationUnfocused broadband (5.4–14.9 MHz, −6 dB) Panametrics XMS-310 contactultrasound transducer (Panametrics Inc.), Aarthroscopic indentation instrument (Artscan 200,Artscan)Tissue modulus, viscoelasticity, US reflection provide a sensitive method for distinction between normal and degenerated cartilage.Laasanen et al. (2003)Mechano-Acoustic Diagnosis of Cartilage Degeneration and RepairB-mode US imaging instrument (DermaScan-C, Cortex Tech), 20 MHz. Arthroscopic indentation instrument (Artscan 200, Artscan)Detailed structural properties of cartilage and subchondral bone could be obtained by B-mode US.Liu & Ebbini (2007)Viscoelastic Property Measurement in Thin Tissue Constructs using UltrasoundDual element US transducer, 5MHz ARF transducer, confocal 25MHz imaging transducer, 200µm diameter, 1-mm depthFrequency response exhibited resonant behaviour determined by stiffness and thickness of tissues.Liu et al. (2008)Rolling Mechanical Imaging: A Novel Approach for Soft Tissue Modelling and Identification during Minimally Invasive SurgeryForce/Torque sensor (ATI NANO 17), 2-mm, 3-mm, 4-mm indentation depthsRolling mechanical image was capable of capturing tissue stiffness, and characterization of force-tissue deflection. Mirbagheri & Dargahi (2005)Finite Element Analysis of a Membrane-Type Piezoelectric Tactile Sensor with Four Sensing ElementsFEM of membrane-type piezoelectric tactile sensor with 4 sensing elements. Ultra-thin circular PVDF, 45 mm radius, 25 µm thickness (see Table 2). Crosstalk, complexity, and fragility are problems solved by the sensor. The sensor exhibited linearity, high sensitivity, and high dynamic range.Najarian et al. (2006)A Novel Method in Measuring the Stiffness of Sensed Objects with Applications for Biomedical Robotic Systems110 µm PVDF film (Goodfellow Co.) between plexiglass bases, Applied force range: 0.01-1NThe sensor exhibited reasonable accuracy (20 percent error) of sensed objects.Najarian et al. (2006)Design and Fabrication of Piezoelectric-based Tactile Sensor for Detecting ComplianceSensor design, 1.44mm diameter PVDF film, modulus of elasticity: 1.8-2.7 GPaAn innovative endoscopic piezoelectric tactile sensor has been designed detecting sensed object compliance.Nieminen et al. (2007)Ultrasound Speed in Articular Cartilage under Mechanical CompressionPulse-echo method. Nonfocused US transducer (V-116, Panametrics), 7.1-14.2 MHz, -3dB. 0.5-100 MHz pulse receiver board for acoustic pulses (PAC-IPR-100, Physical Acoustic Corp.)US speed in articular cartilage is controlled by collagen orientation and void ratio, and depends on strain rate. Compression related errors could in strain and elastic properties need to be eliminated in mechano-acoustic measurements.Nieminen et al. (2006)Acoustic Properties of Articular Cartilage under Mechanical StressPulse-echo method. Non-focused US transducer (VM-116, Panametrics), 7.1-14.2 MHz, -3dB. Acoustic pulses 0.5-100MHz pulse receiver board (PAC-IPR-100, Physical Acoustic Corporation)Mechano-acoustic techniques assuming constant acoustic properties are not significantly impaired by rearrangement of interstitial matrix, especially collagen fibres.Puangmali et al. (2007)State of the Art in Force and Tactile Sensing for Minimally Invasive SurgeryDisplacement-based sensors, current-based sensors, resistive-based sensors, pressure-based sensors, resistive-based sensors, capacitive-based sensors, acoustic-based sensors, and optical-based sensors (see Table 1).Haptic perception enables measurement of tissue hardness, properties, and evaluation of anatomical structures during MIS.Rago et al. (2007)Elastography: New Developments in Ultrasound for Predicting Malignancy in Thyroid NodulesLineartransducer, central frequency of 10 MHz (Hitachi EUB 8500, Hitachi Medical Systems).Elastography is a new technique for estimation of tissue stiffness, and has potential application in diagnosis of thyroid cancer.Riekkinen et al. (2006)Influence of Overlying Soft Tissues on Trabecular Bone Acoustic Measurement at Various Ultrasound FrequenciesPortable US instrument (Opbox-01/100, Optel Ltd), Resolution: 8 bits, Frequency:100 MHz. Pulse receiver bandwith: 0.1-25MHz, -6dB. US transducers (Panametrics)Uncertainties related to in vivo US measurements could be reduced by numerical correction of overlying soft tissue contribution.Riekkinen et al. (2006)Acoustic Properties of Trabecular Bone—Relationships to Tissue CompositionAcoustic measurements. US System (UltraPAC, Physical Acoustic Co), 500MHz A/D-board, 0.2-100MHz pulse receiver. US transducers (Panametrics V304, Panametrics Inc.), 2.25 MHz, Effective range: 1.53-3.8 MHz, -6dBThere were low associations (|r| ≤ 0.4) between composition of calcified matrix and US parameters. Acoustic properties of healthy trabecular bone are affected by content and structure of calcified matrix, instead of composition.Saarakkala et al. (2004)Mechano-Acoustic Determination of Young’s Modulus of Articular CartilageUS transducer mounted on instrument tip (Artscan 200, Artscan) Unfocused miniature 10MHz contact transducer as indenter. Strain gauge inside rod. US speed could predict variation in Young’s modulus and dynamic modulus of cartilage (47 and 53 percent respectively). Dynamic modulus showed significant linear correlation with Young’s modulus and dynamic modulus of cartilage. Acoustic parameters were indicative of articular cartilage mechanical properties.Stallkamp & Schraft (2005)A Technical Challenge for Robot-Assisted Minimally Invasive Surgery: Precision Surgery on Soft TissueDescription of challenges for robot-assisted MIS.Common sensors such as CT, MRI, US, or x-ray cannot acquire extensive data that is accurate and up-to-date for use with a robot.Tavakoli et al. (2005)Haptic Interaction in Robot-Assisted Endoscopic Surgery: A Sensorized End-EffectorPhantom haptic devices ((SensAble Tech. Inc.), strain gaugesLack of haptic feedback is a major drawback in robotic systems. The novel end effector is capable of actuating a tip and interaction with environment.Töyräs et al (2002)Ultrasonic Characterization of Articular CartilageA mode pulse echo measurements. High frequency broadband US instrument (Minihorst Osteoson DCIII, Minihorst), 22Hz mean frequency. High frequency US is a useful tool for the evaluation of articular cartilage, and measurement of thickness of articular cartilage. Töyräs et al. (1999)Characterization of Enzymatically Induced Degradation of Articular Cartilage using High Frequency UltrasoundA-, B-mode pulse echo measurements. High frequency, broadband US instrument (Minohorst Osteoson DCIII, Minihorst) Mean frequency: 22 MHz, Range: 3-35 MHz, - 3dB.High frequency US provides a sensitive technique for analysis of structure and properties of cartilage, for in vivo use.Yanagihara et al. (2008).Development of a Precise Control Method for a Medical Robot Working with Stiff Tissues during Hip-Joint SurgeryPrototype muscle scraping robot, 9 DOF, tension sensors for measurement of load.A prototype muscle scraping robot has been developed.Zbyszewski et al. (2008)Wheel/Tissue Force Interaction: A New Concept for Soft Tissue Diagnosis during MISSpherical sensor, 9 mm spherical ball, optical sensing system Wheel based force sensor, 8 mm diameter aluminium wheel, 8 mm width. Force/torque sensor (ATI NANO17SI-12-0.12) . Mitsubishi manipulator (RV-6SL). Indentation depth: 2-6mmA novel force sensor has been developed for localization of tissue abnormalities, and feedback. Tactile information over large area could be acquired rapidly. Young’s modulus characterizes compressive stiffness of an elastic material. The Young’s modulus could be determined by measuring the speed of sound and density of an isotropic material acoustically. US parameters such as speed and reflection could be used to measure mechanical properties of articular cartilage (Saarakkala et al., 2004). High frequency ultrasound provides a sensitive technique for the analysis of cartilage structure and properties, and could be used in vivo for arthroscopy as a quantitative probe (Toryas et al., 1999). Töyräs et al. (2002) concluded that US was a useful tool for evaluation of articular cartilage, and measurement of cartilage thickness. Riekkinen et al. (2007) used ultrasound measurements to investigate tissue composition and properties on acoustic properties of human trabecular bone at 2.25 MHz, and concluded that acoustic properties of healthy trabecular bone were significantly affected by structure and content of calcified matrix, instead of composition. Quantitative ultrasound measurements include through-transmission and pulse-echo techniques. Attenuation, speed, backscattering, and reflection are parameters of US techniques. Kiviranta et al. (2008) determined that ultrasound reflection measurement could be used for the diagnostics of osteoarthritis, and arthroscopic indentation instruments could be used in the evaluation of cartilage mechanical properties. US coefficient for articular surface was recorded using a device with integrated force gauge and flat miniature transducer (Ф=3mm). 140kPa was applied for determination of tissue thickness, and sequences of short-term and 5% compression were performed. The indentation device enabled measurement of dynamic modulus, thereby enabling quantitative evaluation of cartilage degeneration. Saarakkala et al. (2004) demonstrated that 47 percent and 53 percent of Young’s modulus and dynamic modulus of cartilage could be predicted by sound speed. The reliability of US techniques is affected by soft tissues overlying the bones, in applied frequency range of 0.3-6.7MHz. Riekkinen et al. (2006) introduced numerical correction, a technique for the inclusion specific soft tissue; adipose and lean tissue, acoustic characteristics. Mechanical properties, US parameters and their interrelationships, and errors from soft tissues depend on US frequency. The numerical correction technique resulted in approximately 50 percent reduction in errors from soft tissues. B-mode image involves the presentation of US travel time in two dimensions; US pulse along one axis, and lateral position of transducer along the second axis. Cartilage and subchondral bone structural properties could be represented by B-mode imaging. This provides a means for early degeneration of cartilage diagnosis (Laasanen et al., 2003). Tissue modulus, viscoelasticity and ultrasound reflection could be combined to develop a sensitive method for distinguishing between normal and degenerated cartilage (Laasanen et al., 2002). Hakulinen et al. (2006) studied the relationships between US parameters and microstructure of trabecular bone; and demonstrated that relationships were frequency dependent, and speed showed the strongest association with structural parameters at 5 MHz (center frequency) in comparison to attenuation, reflection and backscatter. US speed in articular cartilage varies depending on loading conditions, contrary to the assumption that US speed is constant under mechanical compression. This limits the application of the technique for determination of elastic properties along the axis of US propagation. Nieminen et al. (2007) developed a linear model to predict the variation of US speed; and suggested that US speed in articular cartilage was controlled by collagen orientation and void ratio, and depended on imposed strain rate. There was a need to eliminate compression-related errors in measured strain. Operating space in orthopaedic surgery is comprised of muscle tissues, tendons, and ligaments. Operating space should be created by scraping and retracting stiff tissues, while not damaging nerves and blood vessels. Yanagihara et al. (2008) developed a muscle scraping robot for surgical applications.ReferencesDargahi, J. & Najarian, S. (2003). An Endoscopic Force-Position Sensor Grasper with Minimum Sensors. Can. J. Elect. Comput. Eng. 28 (3/4), 155-161. Dargahi, J. & Najarian, S. (2005). Advances in Tactile Sensors Design/Manufacturing and Its Impact on Robotics Applications – A Review. Industrial Robot: An International Journal. 32 (3), 268-281. Dargahi, J., Parameswaran, M. & Payandeh, S. (2000). A Micromachined Piezoelectric Tactile Sensor for an Endoscopic Grasper—Theory, Fabrication and Experiments. Journal of Microelectromechanical Systems. 9 (3), 329-335.Gray, B. & Fearing, R. (1996). A Surface Micromachined Microtactile Sensor Array. Proceedings of the 1996 IEEE International Conference on Robotics and Automation. 4 (96), 1-6.Hakulinen, M., Day, J., Toyras, J., Weinans, H. & Jurvelin, J. (2006). Ultrasonic Characterization of Human Trabecular Bone Microstructure. Phys. Med. Biol. . 51 (1), 1633–1648.Harsányi, G. (2001). Sensors in Biomedical Applications. Sensor Review. 21 (4). Hosseini, M., Najarian, S., Motaghinasab, S. & Dargahi, J. (2006). Detection of Tumours Using a Computational Tactile Sensing Approach. Int J Med Robotics Comput Assist Surg. 2 (1), 333–340.Hwang, E., Kim, Y. & Ju, B. (2003). Surface-Micromachined Flexible Polysilicon Sensor Array. IEEE. 0-7803-7744-3 (03), 582-585. Jalkanen, V., Andersson, B., Bergh, A., Ljungberg, B., & Lindahl, O. (2006). Prostate Tissue Stiffness as Measured With a Resonance System: A Study on Silicone and Human Prostate Tissue in Vitro. Med Bio Eng Comput. 44 (1), 593–603.Jalkanen, V., Andersson, B., Bergh, A., Ljungberg, B. & Lindahl, O. (2007). Spatial Variations in Prostate Tissue Histology as Measured By a Tactile Resonance Sensor. Physiol. Meas. 28 (1), 1267–1281. Kiviranta, P., Lammentausta, E., Toyras, J., Kiviranta, I. & Jurvelin, J. (2008). Indentation Diagnostics of Cartilage Degeneration. Osteoarthritis and Cartilage. 16 (1), 796-804. Laasanen, M., Toyras, J., Hirvonen, J., Saarakkala, S., Korhonen, R., Nieminen, M., Kiviranta, I. & Jurvelin, J. (2002). Novel Mechano-Acoustic Technique and Instrument for Diagnosis of Cartilage Degeneration. Physiol. Meas. . 23 (1), 491–503. Laasanen, M., Töyräs, J., Vasara, A., Hyttinen, M., Saarakkala, S., Hirvonen, J., Jurvelin, J. & Kiviranta, I. (2003). Mechano-Acoustic Diagnosis of Cartilage Degeneration and Repair. J Bone Joint Surg Am. 85 (1), 78-84.Liu, D. & Ebbini, E. (2007). Viscoelastic Property Measurement in Thin Tissue Constructs using Ultrasound. IEEE. 1051-0117 (07), 436-439.Liu, H., Noonan, D., Althoefer, K., & Seneviratne, L. (2008). Rolling Mechanical Imaging: A Novel Approach for Soft Tissue Modelling and Identification during Minimally Invasive Surgery. IEEE International Conference on Robotics and Automation. May (19-23), 845-850. Mirbagheri, A. & Dargahi, J. (2005). Finite Element Analysis of a Membrane-Type Piezoelectric Tactile Sensor with Four Sensing Elements. IEEE. 1 (1), 353-356. Najarian, S., Dargahi, J. & Zheng, X. (2006). A Novel Method in Measuring the Stiffness of Sensed Objects with Applications for Biomedical Robotic Systems. Int J Med Robotics Comput Assist Surg. 2 (1), 84–90.Najarian, S., Dargahi, J., Molavi, M. & Singh, H. (2006). Design and Fabrication of Piezoelectric-based Tactile Sensor for Detecting Compliance. IEEE . ISIE (July ), 9-12.Nieminen, H., Julkunen, P, Töyräs, J. & Jurvelin J. (2007). Ultrasound Speed in Articular Cartilage under Mechanical Compression. Ultrasound in Med. & Biol. 33 (11), 1755–1766. Nieminen, H., Töyräs, J., Laasanen, M. & Jurvelin, J. (2006). Acoustic Properties of Articular Cartilage under Mechanical Stress. Biorheology . 43 (1), 523–535. Puangmali, P., Althoefer, K. & Seneviratne, L. (2007). State of the Art in Force and Tactile Sensing for Minimally Invasive Surgery. IEEE. 1-11.Rago, T., Santini, F., Scutari, M. , Pinchera, A. & Vitti, P. (2007). Elastography: New Developments in Ultrasound for Predicting Malignancy in Thyroid Nodules. The Journal of Clinical Endocrinology & Metabolism. 92 (8), 2917–2922. Riekkinen, O., Hakulinen, M., Timonen, M., Töyräs, J., Jurvelin, J. (2006). Influence of Overlying Soft Tissues on Trabecular Bone Acoustic Measurement at Various Ultrasound Frequencies. Ultrasound in Med. & Biol. 22 (7), 1073–1083. Riekkinen, O., Hakulinen, M., Lammi, M., Jurvelin, J., Kallioniemi, A. & Töyräs, J. (2007). Acoustic Properties of Trabecular Bone—Relationships to Tissue Composition . Ultrasound Med Biol. . 33 (9), 1438-44.Saarakkala, S., Korhonen, R., Laasanen, M., Töyräs, J. & Rieppo, J. (2004). Mechano-Acoustic Determination of Young’s Modulus of Articular Cartilage. Biorheology . 41 (1), 167–179. Stallkamp, J. & Schraft, R. (2005). A Technical Challenge for Robot-Assisted Minimally Invasive Surgery: Precision Surgery on Soft Tissue. The International Journal of Medical Robotics and Computer Assisted Surgery. 1 (2), 48-52.Tavakoli, M., Patel, R. & Moallem, M. (2005). Haptic Interaction in Robot-Assisted Endoscopic Surgery: A Sensorized End-Effector. Int J Medical Robotics and Computer Assisted Surgery . 1 (2), 53–63.Töyräs, J., Nieminena, H., Laasanen, M., Nieminenb, M., Korhonen, R., Rieppo, J., Hirvonen, J., Helminenb, H. & Jurvelin, J. (2002). Ultrasonic Characterization of Articular Cartilage. Biorheology. 39 (1), 161–169. Toyras, J., Rieppo, J, Nieminen, M. & Jurvelin, j. (1999). Characterization of Enzymatically Induced Degradation of Articular Cartilage using High Frequency Ultrasound. Phys. Med. Biol.. 44 (1), 2723–2733. Yanagihara, M., Okamoto, J., Fujie, M. & Mitsui, N. (2008). Development of a Precise Control Method for a Medical Robot Working with Stiff Tissues during Hip-Joint Surgery. Proceedings of the 2007 IEEE International Conference on Robotics and Biomimetics. 978-1-4244-1758-2 (08), 342-348. Zbyszewski, D., Liu, H., Puangmali, P., Althoefer, K., Nunes, C., Seneviratne, L., Challacombe, B., Murphy, D. & Dasgupta, P. (2008). Wheel/Tissue Force Interaction: A New Concept for Soft Tissue Diagnosis during MIS. 30th Annual International IEEE EMBS Conference. 978-1-4244-1815-2 (08), 5556-5559.