- Email: [email protected]
A tactile sensor for measuring hardness of soft tissue with applications to minimally invasive surgery
Accepted Manuscript Title: A tactile sensor for measuring hardness of soft tissue with applications to minimally invasive surgery Authors: Lei Zhang, Feng Ju, Yanfei Cao, Yaoyao Wang, Bai Chen PII: DOI: Reference:
S0924-4247(17)31074-9 http://dx.doi.org/10.1016/j.sna.2017.09.012 SNA 10320
To appear in:
Sensors and Actuators A
Received date: Revised date: Accepted date:
6-6-2017 8-9-2017 8-9-2017
Please cite this article as: {http://dx.doi.org/ This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A tactile sensor for measuring hardness of soft tissue with applications to minimally invasive surgery Lei Zhang, Feng Ju, Yanfei Cao, Yaoyao Wang, Bai Chen*
School of Mechanical and Electrical Engineering Nanjing University of Aeronautics and Astronautics No.29, Yudao Street, Nanjing, 210000, Jiangsu Province, China [email protected]
Highlights
Tactile sensor for measuring hardness of soft tissue with application to minimally invasive surgery.
A spiral metal plate is designed to reduce the resonant frequency of the sensor which can restrict the impact brought by the effective mass of tissue.
Only one piece of PZT ceramic is used as both actuator and sensing element, reducing the number of components of the sensor.
Simple structure and small size (8mm×8mm×5mm) suitable for MIS.
Abstract—This paper presents a novel tactile sensor for measuring hardness of soft tissue in minimally invasive surgery (MIS). The proposed tactile sensor consists of a piezoelectric ceramic plate, a spiral metal plate and a probe. The resonant frequency of the sensor shifts when the sensor contact with a tissue. For restricting the impact brought by the effective mass of the tissue, the spiral metal plate is designed to reduce the resonant frequency of the sensor. Another feature of this sensor is that only one piece of lead zirconate titanate (PZT) is used as both actuator and sensing element. So the structure of the sensor is very simple which can be easily miniaturized and is suitable for MIS. The finite element analyses are carried out to verify the feasibility of the sensor and compare with the experimental results. Several silicone samples are used to test the performance of the sensor and the results show the ability of the sensor to measure hardness of soft tissue and detect lumps inside tissue. Keywords—Tactile sensor; Hardness measurement; Biological tissue; Resonant frequency shift; Minimally invasive surgery.
1. Introduction In living organisms, the physical characteristics such as hardness will change when lesions occur in tissues. The hardness of a cancer tumor is significantly higher than normal biological tissue. In traditional surgery, the surgeon can acquire the shape, size and location information of tumors by touching and pressing tissues. However, this can cause large traumas to patients and lengthen the recovery time. Minimally invasive surgery (MIS) becomes more and more popular due to many advantages such as minimized trauma and reduced recovery time [1-2]. However, the surgeon cannot judge the existence of lesions or get the information of tumors in vivo because of the small traumas. Lacking tactile information limits the development of MIS. To solve this problem, several tactile sensors have been developed to provide a sense of touch in MIS. Tactile sensors for measuring hardness can be divided into two categories: time-domain measurements and frequency-domain measurements [3]. In the first category, the contact force between sensor and tissue and the hardness of tissue determine the degree of deformation of the tissue. Larger force and softer tissue lead to greater deformation. Based on this principle, several tactile sensors are presented [4-9]. However, these sensors require significant deformation of the tissue which is harmful to the tissue. Moreover, the measurement results of these sensors can be easily disturbed by many factors such as nonlinear friction, electrical interference and so on. In the second category, the natural frequency of the sensor system shifts when the sensor contacts with the tissue. Tissues with different hardness correspond to different resonant frequencies. Most of these tactile sensors use piezoelectric actuators to drive other components like cantilevers [10-14] or axial rods [15-22] to measure the hardness of the tissue. The resonant frequency of a piezoelectric disk is in the kilohertz range, which is too high for the biological tissues to show viscoelasticity. Also, the measuring results will have a significant deviation from the actual values because the effective mass of the tissue can cause significant frequency shifts under high measuring frequency[23]. The resonant frequency of the whole sensing system can be effectively reduced by combining piezoelectric ceramics with other mechanical elements like springs and masses [3, 23]. These elements form a sensing system, the resonant frequency of the system is measured instead of PZT itself, based on the same principle. The resonant frequency shifts of the whole sensing system correspond to different hardness. This method can reduce the impact that the effective mass of the tissue brings, however, it also leads to complex structure of the sensor, making it hard to be miniaturized and used in MIS. The proposed tactile sensor in this paper is based on the frequency-domain measurement. Only one piece of PZT is used as both actuator and sensing element, combined with a piece of spiral metal plate and a probe. Alternating voltage signal is applied to the PZT to drive the sensing system. The electrical impedance of the PZT is measured, which can reflect the impedance of the sensing system. Abrupt change of the system’s impedance corresponds to the resonant frequency, which is related to the hardness of tissue. The resonant frequency of the whole sensing system is reduced by the spiral metal plate (about 230Hz), which is far lower than that of the PZT itself (about 721kHz). Under this frequency, the impact brought by the effective mass can be restricted. The structure of the proposed tactile sensor is very simple which can be easily miniaturized and is suitable for MIS. The finite element method (FEM) including static structural analysis, modal analysis and harmonic response analysis are carried out to verify the feasibility of the sensor and compare with the experimental results. Several silicone samples are used to test the sensor, the results show the sensor’s ability of measuring hardness and detecting lump inside the tissue.
2. Structure and principle 2.1 Structure of the tactile sensor The proposed tactile sensor consists of a piece of PZT plate, a piece of spiral metal plate and a cylindrical probe, the concept structure is shown in Fig. 1. The PZT bonded on the upper side of the spiral metal plate has a cuboid structure. D31 polarization is used in the PZT which means vibration is generated in the length direction while electric filed is applied to the thickness direction. The spiral metal plate with thickness of 0.1mm is fabricated from a complete metal plate by using linear cutting machine. The probe is a slender cylinder with a spherical tip, bonded on the other side of the spiral metal plate to touch the tissue. The spherical tip can protect the tissue from scratches. These three parts form a mechanical system for measuring the hardness of tissue. The special structure of the spiral metal plate ensures the low resonant frequency of this system by constraining the terminal surface of the PZT plate, as shown in Fig. 1. This constraint turns the mechanical system to act like a cantilever beam so that the resonant frequency of this system is much lower than that of the PZT itself. 2.2 Basic principle The basic principle of this tactile sensor can be explained in Fig. 2. The Kelvin viscoelastic model consists of a spring and damping is often used to describe the biological tissues. Here, one Kelvin viscoelastic model is enough to explain the principle. In the model, the force 𝐹 is generated by the PZT,
k1
and c 1 are the equivalent stiffness and damping of the measuring mechanical
system respectively, 𝑚𝑡𝑖𝑝 is the mass of the tip with displacement𝑥. Similarly, 𝑘𝑚 and𝑐𝑚 are the equivalent stiffness and damping of the tissue respectively,
mm
is the effective mass of the tissue. The harmonic excitation is provided by the piezoelectric actuator,
then the force and displacement are transmitted to the indenter head and resonance occurs to the measuring system. The state-space model in this case is given as follows: 𝑥̇ =𝐴𝑥 + 𝐵
(1)
where 𝑥 = (𝑥 𝑥̇ )𝑇 0 𝑘 + 𝐴 = [ 1 𝑘𝑚 −𝑚0 𝐵 = (0
1 𝑐1 + 𝑐𝑚 ] −𝑚0 𝐹 𝑇 ) 𝑚0
𝑚0 = 𝑚𝑡𝑖𝑝 + 𝑚𝑚 and the resonant frequency of the whole sensing system is: 𝜔𝑛 = √
𝑘1 +𝑘𝑚
(2)
𝑚𝑡𝑖𝑝 +𝑚𝑚
The impedance of the tissue 𝑍𝑚 can be expressed as follow: 𝑍𝑚 = 𝑐𝑚 + 𝑗𝜔𝑚𝑚 +
𝑘𝑚 𝑗𝜔
(3)
The expressions of the components of 𝑍𝑚 can be seen in appendix. As can be seen from Eq. 2, the resonant frequency shifts when the sensor touches the tissue because of the mechanical parameters of the tissue. An important parameter is the stiffness k m , which has a significant influence on the resonant frequency. When k m increases, the resonant frequency 𝜔𝑛 increases but the rising speed of 𝜔𝑛 gets smaller. That is why tissues with different hardness correspond to different resonant frequencies. Hardness
changes when lesions occur in tissues, so judging the existence of a lesion or detecting tumors in vivo by measuring the shifts of resonant frequency is feasible. The effective mass of the tissue
mm
is also a factor affecting the resonant frequency, Eq. 3 shows
that the dominant factor affecting the impedance is the effective mass dominant factor is
km
mm
when the excitation frequency ω is high, and the
when ω is low. Reducing the frequency to restrict the impact brought by the effective mass is necessary
as mentioned above. As shown in Fig. 3, an alternating voltage is applied on the electrodes of the PZT, leads to the extension or shortening of the PZT in the length direction. Then the deformation in PZT transmits to the spiral metal plate and the probe, and leads to the resonance of the measuring system and the minimum value of its mechanical impedance. Here, only one PZT is used as both actuator and sensing element. The electrical impedance of the PZT can reflect the mechanical impedance of the sensing system. The resonant frequency of the tactile sensor shifts when the probe touches the tissue, then the hardness of the tissue can be detected by analyzing the collected current signals. 3. FEM analyses After presenting the structure and principle of the tactile sensor, this section is devoted to accomplish the static structural, modal and harmonic response analyses. The materials of each component are listed in Table 1. The measurement of hardness is divided into two steps. First, a contact force is generated between the probe and tissue. Then, sweep voltage signal is applied to the PZT and the resonant frequency will be detected by the computer through the feedback signal. For corresponding to the actual situation, the static structural analysis is finished first to simulate the first step. The modal analysis is used to determine the natural frequency and sweep range. Finally, the harmonic response analysis is carried out to get the simulation curve of the electrical impedance based on the result of the static structural analysis. 3.1 Static structural analysis In this simulation, a simple spiral metal plate is designed, the outer size of the metal plate is 8mm×8mm, with the thickness of 0.1mm. The size of the PZT is 7mm×1mm×0.5mm. The probe is 4.5mm in length and 1mm in diameter. A cube below, contacting with the tip of the probe, is the tissue. The boundary conditions are set as follows: The bottom of the tissue is fixed; a force of 30mN, along the Z axis, is applied to the side face of the PZT and the displacement of this face is also limited along the Z axis. When the PZT and the probe move down, the spiral metal plate will deform under such constraints. However, the displacement of the probe is less than the PZT because of the reaction of the tissue. Under the same force, the softer the tissue, the greater the deformation of the tissue, which can be seen in Fig. 4. Four samples with elasticity modulus of 100kPa, 200kPa, 300kPa and 400kPa are used to simulate the deformation. The indentation of tissue gets shallower and shallower as predicted above. The actual contact is well simulated by result of the static structural analysis, which is the basis of the following coupling analysis and the comparison of the experimental results.Modal analysis To evaluate the natural frequencies and the mode shapes of the tactile sensor, a modal analysis is performed, the results are shown in Table 2 and Fig. 5.
The only constraint in the modal analysis is the fixation of the side face of the PZT, as shown in Fig. 5 (a). The first natural frequency is 230.02Hz, far less than the natural frequency of the PZT. At such frequency, the impact brought by the effective mass and the low-frequency perturbation can be reduced, which means the first natural frequency is suitable for measuring the hardness of biological tissue. The other three frequencies are obviously larger than 230.02Hz, and the deformation of the probe is offset from
the vertical direction (Z axis), which cannot be picked. The result of the modal analysis can be used as a reference for the harmonic response analysis. 3.2 Harmonic response analysis The harmonic response analysis, taking into account the result of the static structural analysis, is carried out, in order to obtain the electrical impedance of the PZT when the tactile sensor touches the tissue. The material of the ceramic is PZT-5A, whose physical parameters can be find in appendix. Four silicone specimens with different hardness values are tested, as shown in Table 3. An alternating voltage with varying frequency (200 to 300Hz) is applied on the polarization surface of the PZT. The frequency range is chosen by referring to the results of the modal analysis, where the tactile sensor will resonate near the first natural frequency. The electrical impedance signals corresponding to the four silicone samples can be obtained by solving the harmonic response analysis, as shown in Fig. 6. Abrupt change of the amplitude of each signal occurs at a particular frequency, which is exactly the resonant frequency. The resonant frequencies corresponding to four samples of different hardness are 205Hz, 224Hz, 233Hz and 246Hz, respectively. The significant differences between the four frequencies show the sensor’s ability of measuring hardness, which verifies the feasibility of the tactile sensor. Then, 14 silicone samples of different hardness values are analyzed in order to calibrate the sensor, Table 4 and Fig. 7 shows the hardness of the 14 samples and the calibration curve respectively. The curve is fitted in MATLAB, the formula of fitting curve is: 𝑓𝑠 (𝑥) = −3435𝑥 −0.8432 + 254.3
(4)
where 𝑥 stands for the elasticity modulus in kPa, and 𝑓𝑠 (𝑥) stands for the simulating resonant frequency in Hz. The coefficient of determination of the fitted curve is 99.89%. The first half of the curve is steeper, which means the tactile sensor is sensitive to the tissue whose elasticity modulus is less than 1.5MPa. This range of elasticity modulus is consistent with that of most biological soft tissues, which means the proposed tactile sensor is suitable for the hardness measurement of biological tissue. 4. Experiments and results For further verification of the sensor’s feasibility, a prototype of the sensor is fabricated and an experimental system is set up, as shown in Fig. 8 and Fig. 9. A data acquisition card (National Instruments, NI USB-7856R) connecting to a computer is used as a signal generator and a data collector, controlled by a data acquisition and analysis software based on LabVIEW. The tactile sensor is fixed to the linear stage by a fixture. An oscilloscope is used to show the feedback signal more intuitively and the electronic balance is used to control the contact force. The measurement process can be divided into the following steps: First, a contact force of 30mN is generated between the sensor and experimental samples. Then, an alternating voltage signal is applied to the PZT and the current flowing through is used as the feedback signal, measured by an amplifying circuit and acquired by the data acquisition card. After that, the electrical impedance of the PZT can be calculated from the voltage and current signal, which is related to the mechanical impedance of the whole sensing system. Finally, the resonant frequency can be extracted from the electrical impedance, which reflects the information of hardness. 4.1 Hardness measurement The performance of the tactile sensor is evaluated on five silicone samples, whose hardness changes according to concentration. Fig. 10 and Table 5 show the five silicone samples and the hardness of them respectively.
In order to compare with the simulation results, the sensor is fixed to the linear stage, similar to the constraint in simulation. The natural frequency of the sensor itself is measured first, which is 232.3Hz, very close to the result of the modal analysis, which is 230.02Hz. Then, the contact force between the sensor and the sample is controlled by the displacement of the linear stage. As 30mN is set in the static structural analysis, the displacement is controlled to make sure the actual contact force is 30mN, which can be shown in the electronic balance. Each sample is measured 20 times and the results are shown in Fig. 11, comparing with the simulation fitting curve. The experimental results show the sensor performs well in hardness measurement. The data are in good agreement with the simulation curve. A small deviation in higher modulus range is observed between the FEM simulation result and the experimental result. This is due to the differences between the actual condition and the simulation setting, such as the material characteristics, the manufacturing errors and meshing accuracy. To meet the actual measurement results, the calibration curve of the sensor is corrected by the experimental data, as shown in the following formula: 𝑓𝑒 (𝑥) = −2305𝑥 −0.7243 + 266.6
(5)
where 𝑥 stands for the elasticity modulus in kPa, and 𝑓𝑒 (𝑥) stands for the experimental resonant frequency in Hz. The coefficient of determination of the experimental fitting curve is 99.91%. In this paper, the PZT used in the prototype of the sensor is a general product, whose size is small and the thickness is large, leading to the large impedance of PZT itself and the small amplitude of the feedback current signal. So the feedback signal can be easily affected by the noise signal of the sensing system, which reduces the reliable measuring range of the sensor. As to the sensor presented in this paper, the actual reliable measuring range of elasticity modulus is smaller than 500kPa, which is narrower than the simulating measurement range because of the impact brought by the noise. However, the accuracy of the measurement can be improved and the reliable measuring range of the sensor can be expanded by increasing the signal-to-noise ratio. In the next step of this research, the PZT will be made by customization to enlarge its size and reduce the thickness to reduce its impedance. A filter and a signal amplifier will be added to the sensing system. 4.2 Lump detection After verifying the sensor’s ability of measuring hardness, another silicone sample with a lump inside is made to test whether the lump can be detected by the sensor. As shown in Fig. 12(a), the silicone sample is 30mm in diameter and 15mm thick, a coin is used as the lump locating about 1mm below the surface of the silicone sample.
In order to compare with the experimental fitting curve, a force of 30mN is applied to the sample and several points of each area are tested. The average of the results is marked in the Fig. 12(b), which is 206.3Hz for the area without lump and 223.5Hz for the area with lump. Each area is marked with red for different depths corresponding to the frequency values, the higher the value, the deeper the color. The elasticity modulus of each area can be calculate referring to Eq. 5, which is 151.8kPa and 246.3kPa respectively. The results show the sensor’s ability of lump detection and the deeper lump can be detected if the contact force becomes larger. 5. Conclusion A new tactile sensor for measuring the hardness of soft tissue is presented. The proposed tactile sensor consists of a piezoelectric ceramic plate, a spiral metal plate and a probe. Only one piece of PZT is used as both actuator and sensing element, which reduces
the amount of sensor’s components. A simple spiral metal plate is designed to reduce the resonant frequency of the sensor for restricting the impact brought by the effective mass of the tissue. The structure of the sensor is very simple, which can be easily miniaturized. The concept structure of the sensor is proposed and working principle is explained. The relationship between resonant frequency and hardness of tissue is discussed by using Kelvin viscoelastic model. FEM analysis including static structural analysis, modal analysis and harmonic response analysis are carried out to verify the feasibility of the sensor and compare with the experimental results. The contact model of the sensor and tissue is obtained by static structural, which is also the basic of the harmonic response analysis. The first four resonant frequencies are solved by the modal analysis, which is the reference for selecting frequency sweep range. The sensor’s ability of measuring hardness is verified and the simulation fitting curve is obtained by harmonic response analysis. A prototype of the sensor is fabricated and the experimental system is established, five silicone samples are used to test the sensor. The feasibility of the sensor is further verified by the experimental results, which has a good agreement with the simulation results. The simulation fitting curve is also corrected by the experimental results. Furthermore, the ability of lump detection is demonstrated by a silicone sample with a lump inside. The application and evolution of this tactile sensor will be helpful for the development of MIS. Acknowledgment This work is supported by the National Natural Science Foundation of China (No.51575256) and the Starting Fund of Nanjing University of Aeronautics and Astronautics (No. 90YAH16062). REFERENCES [1]
Bicchi A, Canepa G, De Rossi D, et al. A sensor-based minimally invasive surgery tool for detecting tissue elastic properties[C]// IEEE International Conference on Robotics and Automation, 1996. Proceedings. IEEE, 1996:884-888 vol.1.
[2]
Dargahi J, Parameswaran M, Payandeh S. A micromachined piezoelectric tactile sensor for use in endoscopic graspers.[J]. Microelectromechanical Systems Journal of, 2000, 9(3):329-335.
[3]
Araghi M H, Salisbury S P. Improved Evaluation of Dynamic Mechanical Properties of Soft Materials With Applications to Minimally Invasive Surgery[J]. IEEE/ASME Transactions on Mechatronics, 2013, 18(3):973-980.
[4]
Eltaib M E H, Hewit J R. Tactile sensing technology for minimal access surgery––a review[J]. Mechatronics, 2003, 13(10):1163-1177.
[5]
Kalanovic D, Ottensmeyer M P, Gross J, et al. Independent testing of soft tissue visco-elasticity using indentation and rotary shear deformations.[J]. Studies in Health Technology & Informatics, 2003, 94:137-143.
[6]
Liu H, Noonan D P, Challacombe B J, et al. Rolling Mechanical Imaging for Tissue Abnormality Localization During Minimally Invasive Surgery[J]. IEEE Transactions on Biomedical Engineering, 2010, 57(2):404-414.
[7]
Liu H, Li J, Song X, et al. Rolling Indentation Probe for Tissue Abnormality Identification During Minimally Invasive Surgery[J]. IEEE Transactions on Robotics, 2011, 27(3):450-460.
[8]
Kalantari M, Shen J J, Dargahi J, et al. Localization of annulus with a tactile sensor[C]// IEEE, Northeast Bioengineering Conference. IEEE, 2011:1-2.
[9]
Li C G, Shen J J. Design and analysis of a tactile sensor used in minimally invasive surgery[C]// Ieee/asme International Conference on Advanced Intelligent Mechatronics. IEEE, 2013:1454-1457.
[10] Kanda T, Morita T, Kurosawa M K, et al. A flat type touch probe sensor using PZT thin film vibrator[J]. Sensors & Actuators A Physical, 2000, 83(1–3):6775. [11] Barthod C, Teisseyre Y, Géhin C, et al. Resonant force sensor using a PLL electronic[J]. Sensors & Actuators A Physical, 2003, 104(2):143-150. [12] Beekmans S V, Iannuzzi D. Characterizing tissue stiffness at the tip of a rigid needle using an opto-mechanical force sensor[J]. Biomedical Microdevices, 2016, 18(1):1-8.
[13] Valtorta D, Mazza E. Dynamic Measurements of Soft Tissue Viscoelastic Properties with a Torsional Resonator Device[M]// Medical Image Computing and Computer-Assisted Intervention – MICCAI 2004. Springer Berlin Heidelberg, 2004:481-490. [14] Ju F, Ling S F. A micro whisker transducer with sensorless mechanical impedance detection capability for fluid and tactile sensing in space-limited applications[J]. Sensors & Actuators A Physical, 2015, 234:104-112. [15] Omata S, Terunuma Y. New tactile sensor like the human hand and its applications [J]. Sensors & Actuators A Physical, 1992, 35(1):9-15. [16] Murayama Y, Haruta M, Hatakeyama Y, et al. Development of a new instrument for examination of stiffness in the breast using haptic sensor technology[J]. Sensors & Actuators A Physical, 2008, 143(2):430-438. [17] Omata S, Murayama Y, Constantinou C E. Real time robotic tactile sensor system for the determination of the physical properties of biomaterials[J]. Sensors & Actuators A Physical, 2004, 112(2):278-285. [18] Jalkanen V, Lindahl O A. Hand-held resonance sensor instrument for soft tissue stiffness measurements - a first study on biological tissue in vitro[M]// XII Mediterranean Conference on Medical and Biological Engineering and Computing 2010. Springer Berlin Heidelberg, 2010:463-466. [19] Åstrand A P, Jalkanen V, Andersson B M, et al. A Flexible Sensor System Using Resonance Technology for Soft Tissue Stiffness Measurements –Evaluation on Silicone[M]// 15th Nordic-Baltic Conference on Biomedical Engineering and Medical Physics (NBC 2011). Springer Berlin Heidelberg, 2011:21-24. [20] Jalkanen V, Andersson B M, Lindahl O A. Stiffness of a small tissue phantom measured by a tactile resonance sensor[M]// XII Mediterranean Conference on Medical and Biological Engineering and Computing 2010. Springer Berlin Heidelberg, 2010:395-398. [21] Åstrand A P, Andersson B M, Jalkanen V, et al. Initial Measurements on Whole Human Prostate ex vivo with a Tactile Resonance Sensor in Order to Detect Prostate Cancer[J]. Ifmbe Proceedings, 2015, 48:120-123. [22] Åstrand A P, Jalkanen V, Andersson B M, et al. Stiffness measurements on spherical surfaces of prostate models, using a resonance sensor[M]// World Congress on Medical Physics and Biomedical Engineering May 26-31, 2012, Beijing, China. Springer Berlin Heidelberg, 2013:1401-1404. [23] Dhar P R, Zu J W. Design of a resonator device for in vivo measurement of regional tissue viscoelasticity[J]. Sensors & Actuators A Physical, 2007, 133(1):4554.
Biographies Lei Zhang received the Bachelor’s degree from Nanjing University of Aeronautics and Astronautics, Nanjing, China in 2015. He is currently a Master’s degree candidate at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include tactile sensors and surgical robots. Feng Ju received the PhD degree from Nanyang Technological University, Singapore. He is currently an Associate Professor in the School of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics. His research interests include piezoelectric sensors and actuators, robotic tactile sensing and motion control for robotics. Yanfei Cao received the Bachelor’s degree from Nanjing University of Aeronautics and Astronautics, Nanjing, China in 2016. He is currently a Master’s degree candidate at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include surgical robots and cable-driven robots. Yaoyao Wang received the Bachelor’s degree from Southeast University, Nanjing, China in 2011 and the PhD degree from Zhejiang University, Hangzhou, China in 2016.He is currently a Lecturer in the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include robust control and adaptive control of underwater vehicle-manipulator system, design and control of cable-driven manipulators. Bai Chen received the Bachelor’s and PhD degrees from Zhejiang University, Hangzhou, China, in 2000 and 2005, respectively. He is currently a Professor and PhD candidate supervisor at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include design and control of surgical robots, cable-driven robots and sperm-like swimming micro robots.
Spiral metal plate
PZT
Constraint surface
Probe
Fig. 1. Concept structure of the tactile sensor
Piezoelectric Actuator
k1
F
c1 Indenter Head
mtip mm
x km
cm
Tissue
Fig. 2. Connect model of the tactile and tissue
PZT Metal plate
Probe
Fig. 3.
Deformation of the tactile sensor under harmonic excitation
(a)
(b)
(c)
(d) Fig. 4. Results of the static structural analysis: (a) sample with elasticity modulus of 100kPa, (b) sample with elasticity modulus of 200kPa, (c) sample with elasticity
modulus of 300kPa, (d) sample with elasticity modulus of 400kPa
(a)
(b)
(c)
(d) Fig. 5. First four resonant modes: (a) the first mode, (b) the second mode, (c) the third mode, (d) the forth mode
Fig. 6.
Electrical impedance signals corresponding to the four silicone samples
Fig. 7.
The fitting curve fitted by 14 simulating silicone samples
PZT
Probe
Spiral metal plate Fig. 8.
Signal wire 1cm
The prototype of the sensor comparing with a scale
Fig. 9.
The experimental system
Sample# A
Sample# B
Sample# D
Sample# C
Sample# E
Fig. 10. Five silicone samples with different hardness values
Fig. 11. Experimental results comparing with simulation fitting curve
206.3Hz
Points detected
Lump
Silicone sample 223.5Hz
(a)
(b) Fig. 12. Lump detection: (a) the silicone sample with lump inside, (b)
the detecting result
TABLE. 1.
TABLE. 2.
TABLE. 3.
Material of each component
Component
Material
PZT
PZT-5A
Spiral metal plate
Stainless steel
Probe
Stainless steel
The first four resonant frequencies
Mode
Frequency (Hz)
1
230.02
2
326.31
3
390.96
4
1331.9
Four silicone specimens
Code
Shore hardness (HA)
Elasticity modulus (kPa)
A
0
157.5
B
5
279
C
10
414
D
30
1146
TABLE. 4.
Elasticity modulus of 14 silicone samples
Code
TABLE. 5.
Elasticity
modulus
(kPa)
Code
Elasticity
modulus
(kPa)
1
100
8
800
2
200
9
900
3
300
10
1000
4
400
11
1500
5
500
12
2000
6
600
13
2500
7
700
14
3000
Five silicone samples for experimental study
Code
Shore hardness (HA)
Elasticity modulus (kPa)
A
0
157.5
B
5
279
C
10
414
D
30
1146
E
50
2465
APPENDIX 1. The expressions of the components of tissue’s impedance When the sensor presented in this paper is excited to vibrate at its natural frequency and pressed to the tissue, as shown in the Fig. 13, the impedance of the tissue can be expressed as:
𝑍𝑚 = 𝑐𝑚 + 𝑗𝜔𝑚𝑚 +
𝑘𝑚 𝑗𝜔
where 𝑐𝑚 , 𝑚𝑚 and 𝑘𝑚 are equivalent damping, equivalent mass and equivalent stiffness respectively. They can be expressed as:
𝑐𝑚 =
𝐶0 𝜌𝑚 𝑣𝑚 𝑆𝑐 1 − 𝜐𝑚 3
0.1𝜌𝑚 𝑆𝑐2 𝑚𝑚 = 1 − 𝜐𝑚 {
𝑘𝑚 =
2𝐸𝑚 √𝑆𝑐 2) √𝜋(1 − 𝜐𝑚
where 𝐶0 is a function related to the Poisson ratio 𝜐𝑚 , which can be seen as a constant. 𝜌𝑚 is the density of the tissue, 𝑆𝑐 =π𝑎02 is the contact area, 𝑎0 is the radius of the contact area, 𝐸𝑚 is the elasticity modulus of the tissue and 𝑣𝑚 = √ velocity of the wave spread in the tissue.
Z1
Sensor
T issue
Zm
2
a
0
E ffec tiv e mass
Fig. 13 Contact modal of the measurement
2. The material parameters of PZT-5A The material of the PZT set in FEM analyses is PZT-5A, whose physical parameters are: 𝑑=[
0 0 −5.35
0 0 −5.35
0 0 15.78
0 12.26 0
12.26 0 0
0 0] × 10−10 𝐶/𝑁 0
𝐸𝑚 2(1+𝜐𝑚 )𝜌𝑚
is the
12.04 7.52 7.51 𝐸 𝑐 = 0 0 [ 0
7.52 12.04 7.51 0 0 0
7.51 7.51 11.09 0 0 0
8.13 𝜀𝑆 = [ 0 0
0 8.13 0
0 0 0 2.26 0 0
0 0 0 0 2.11 0
0 0 0 × 1010 𝑁/𝑚2 0 0 2.11]
0 0 ] × 10−9 𝐹/𝑚 7.31
where 𝑑, 𝑐 𝐸 , and 𝜀 𝑆 are the piezoelectric matrix, the stiffness matrix and the dielectric matrix, respectively.
S0924-4247(17)31074-9 http://dx.doi.org/10.1016/j.sna.2017.09.012 SNA 10320
To appear in:
Sensors and Actuators A
Received date: Revised date: Accepted date:
6-6-2017 8-9-2017 8-9-2017
Please cite this article as: {http://dx.doi.org/ This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
A tactile sensor for measuring hardness of soft tissue with applications to minimally invasive surgery Lei Zhang, Feng Ju, Yanfei Cao, Yaoyao Wang, Bai Chen*
School of Mechanical and Electrical Engineering Nanjing University of Aeronautics and Astronautics No.29, Yudao Street, Nanjing, 210000, Jiangsu Province, China [email protected]
Highlights
Tactile sensor for measuring hardness of soft tissue with application to minimally invasive surgery.
A spiral metal plate is designed to reduce the resonant frequency of the sensor which can restrict the impact brought by the effective mass of tissue.
Only one piece of PZT ceramic is used as both actuator and sensing element, reducing the number of components of the sensor.
Simple structure and small size (8mm×8mm×5mm) suitable for MIS.
Abstract—This paper presents a novel tactile sensor for measuring hardness of soft tissue in minimally invasive surgery (MIS). The proposed tactile sensor consists of a piezoelectric ceramic plate, a spiral metal plate and a probe. The resonant frequency of the sensor shifts when the sensor contact with a tissue. For restricting the impact brought by the effective mass of the tissue, the spiral metal plate is designed to reduce the resonant frequency of the sensor. Another feature of this sensor is that only one piece of lead zirconate titanate (PZT) is used as both actuator and sensing element. So the structure of the sensor is very simple which can be easily miniaturized and is suitable for MIS. The finite element analyses are carried out to verify the feasibility of the sensor and compare with the experimental results. Several silicone samples are used to test the performance of the sensor and the results show the ability of the sensor to measure hardness of soft tissue and detect lumps inside tissue. Keywords—Tactile sensor; Hardness measurement; Biological tissue; Resonant frequency shift; Minimally invasive surgery.
1. Introduction In living organisms, the physical characteristics such as hardness will change when lesions occur in tissues. The hardness of a cancer tumor is significantly higher than normal biological tissue. In traditional surgery, the surgeon can acquire the shape, size and location information of tumors by touching and pressing tissues. However, this can cause large traumas to patients and lengthen the recovery time. Minimally invasive surgery (MIS) becomes more and more popular due to many advantages such as minimized trauma and reduced recovery time [1-2]. However, the surgeon cannot judge the existence of lesions or get the information of tumors in vivo because of the small traumas. Lacking tactile information limits the development of MIS. To solve this problem, several tactile sensors have been developed to provide a sense of touch in MIS. Tactile sensors for measuring hardness can be divided into two categories: time-domain measurements and frequency-domain measurements [3]. In the first category, the contact force between sensor and tissue and the hardness of tissue determine the degree of deformation of the tissue. Larger force and softer tissue lead to greater deformation. Based on this principle, several tactile sensors are presented [4-9]. However, these sensors require significant deformation of the tissue which is harmful to the tissue. Moreover, the measurement results of these sensors can be easily disturbed by many factors such as nonlinear friction, electrical interference and so on. In the second category, the natural frequency of the sensor system shifts when the sensor contacts with the tissue. Tissues with different hardness correspond to different resonant frequencies. Most of these tactile sensors use piezoelectric actuators to drive other components like cantilevers [10-14] or axial rods [15-22] to measure the hardness of the tissue. The resonant frequency of a piezoelectric disk is in the kilohertz range, which is too high for the biological tissues to show viscoelasticity. Also, the measuring results will have a significant deviation from the actual values because the effective mass of the tissue can cause significant frequency shifts under high measuring frequency[23]. The resonant frequency of the whole sensing system can be effectively reduced by combining piezoelectric ceramics with other mechanical elements like springs and masses [3, 23]. These elements form a sensing system, the resonant frequency of the system is measured instead of PZT itself, based on the same principle. The resonant frequency shifts of the whole sensing system correspond to different hardness. This method can reduce the impact that the effective mass of the tissue brings, however, it also leads to complex structure of the sensor, making it hard to be miniaturized and used in MIS. The proposed tactile sensor in this paper is based on the frequency-domain measurement. Only one piece of PZT is used as both actuator and sensing element, combined with a piece of spiral metal plate and a probe. Alternating voltage signal is applied to the PZT to drive the sensing system. The electrical impedance of the PZT is measured, which can reflect the impedance of the sensing system. Abrupt change of the system’s impedance corresponds to the resonant frequency, which is related to the hardness of tissue. The resonant frequency of the whole sensing system is reduced by the spiral metal plate (about 230Hz), which is far lower than that of the PZT itself (about 721kHz). Under this frequency, the impact brought by the effective mass can be restricted. The structure of the proposed tactile sensor is very simple which can be easily miniaturized and is suitable for MIS. The finite element method (FEM) including static structural analysis, modal analysis and harmonic response analysis are carried out to verify the feasibility of the sensor and compare with the experimental results. Several silicone samples are used to test the sensor, the results show the sensor’s ability of measuring hardness and detecting lump inside the tissue.
2. Structure and principle 2.1 Structure of the tactile sensor The proposed tactile sensor consists of a piece of PZT plate, a piece of spiral metal plate and a cylindrical probe, the concept structure is shown in Fig. 1. The PZT bonded on the upper side of the spiral metal plate has a cuboid structure. D31 polarization is used in the PZT which means vibration is generated in the length direction while electric filed is applied to the thickness direction. The spiral metal plate with thickness of 0.1mm is fabricated from a complete metal plate by using linear cutting machine. The probe is a slender cylinder with a spherical tip, bonded on the other side of the spiral metal plate to touch the tissue. The spherical tip can protect the tissue from scratches. These three parts form a mechanical system for measuring the hardness of tissue. The special structure of the spiral metal plate ensures the low resonant frequency of this system by constraining the terminal surface of the PZT plate, as shown in Fig. 1. This constraint turns the mechanical system to act like a cantilever beam so that the resonant frequency of this system is much lower than that of the PZT itself. 2.2 Basic principle The basic principle of this tactile sensor can be explained in Fig. 2. The Kelvin viscoelastic model consists of a spring and damping is often used to describe the biological tissues. Here, one Kelvin viscoelastic model is enough to explain the principle. In the model, the force 𝐹 is generated by the PZT,
k1
and c 1 are the equivalent stiffness and damping of the measuring mechanical
system respectively, 𝑚𝑡𝑖𝑝 is the mass of the tip with displacement𝑥. Similarly, 𝑘𝑚 and𝑐𝑚 are the equivalent stiffness and damping of the tissue respectively,
mm
is the effective mass of the tissue. The harmonic excitation is provided by the piezoelectric actuator,
then the force and displacement are transmitted to the indenter head and resonance occurs to the measuring system. The state-space model in this case is given as follows: 𝑥̇ =𝐴𝑥 + 𝐵
(1)
where 𝑥 = (𝑥 𝑥̇ )𝑇 0 𝑘 + 𝐴 = [ 1 𝑘𝑚 −𝑚0 𝐵 = (0
1 𝑐1 + 𝑐𝑚 ] −𝑚0 𝐹 𝑇 ) 𝑚0
𝑚0 = 𝑚𝑡𝑖𝑝 + 𝑚𝑚 and the resonant frequency of the whole sensing system is: 𝜔𝑛 = √
𝑘1 +𝑘𝑚
(2)
𝑚𝑡𝑖𝑝 +𝑚𝑚
The impedance of the tissue 𝑍𝑚 can be expressed as follow: 𝑍𝑚 = 𝑐𝑚 + 𝑗𝜔𝑚𝑚 +
𝑘𝑚 𝑗𝜔
(3)
The expressions of the components of 𝑍𝑚 can be seen in appendix. As can be seen from Eq. 2, the resonant frequency shifts when the sensor touches the tissue because of the mechanical parameters of the tissue. An important parameter is the stiffness k m , which has a significant influence on the resonant frequency. When k m increases, the resonant frequency 𝜔𝑛 increases but the rising speed of 𝜔𝑛 gets smaller. That is why tissues with different hardness correspond to different resonant frequencies. Hardness
changes when lesions occur in tissues, so judging the existence of a lesion or detecting tumors in vivo by measuring the shifts of resonant frequency is feasible. The effective mass of the tissue
mm
is also a factor affecting the resonant frequency, Eq. 3 shows
that the dominant factor affecting the impedance is the effective mass dominant factor is
km
mm
when the excitation frequency ω is high, and the
when ω is low. Reducing the frequency to restrict the impact brought by the effective mass is necessary
as mentioned above. As shown in Fig. 3, an alternating voltage is applied on the electrodes of the PZT, leads to the extension or shortening of the PZT in the length direction. Then the deformation in PZT transmits to the spiral metal plate and the probe, and leads to the resonance of the measuring system and the minimum value of its mechanical impedance. Here, only one PZT is used as both actuator and sensing element. The electrical impedance of the PZT can reflect the mechanical impedance of the sensing system. The resonant frequency of the tactile sensor shifts when the probe touches the tissue, then the hardness of the tissue can be detected by analyzing the collected current signals. 3. FEM analyses After presenting the structure and principle of the tactile sensor, this section is devoted to accomplish the static structural, modal and harmonic response analyses. The materials of each component are listed in Table 1. The measurement of hardness is divided into two steps. First, a contact force is generated between the probe and tissue. Then, sweep voltage signal is applied to the PZT and the resonant frequency will be detected by the computer through the feedback signal. For corresponding to the actual situation, the static structural analysis is finished first to simulate the first step. The modal analysis is used to determine the natural frequency and sweep range. Finally, the harmonic response analysis is carried out to get the simulation curve of the electrical impedance based on the result of the static structural analysis. 3.1 Static structural analysis In this simulation, a simple spiral metal plate is designed, the outer size of the metal plate is 8mm×8mm, with the thickness of 0.1mm. The size of the PZT is 7mm×1mm×0.5mm. The probe is 4.5mm in length and 1mm in diameter. A cube below, contacting with the tip of the probe, is the tissue. The boundary conditions are set as follows: The bottom of the tissue is fixed; a force of 30mN, along the Z axis, is applied to the side face of the PZT and the displacement of this face is also limited along the Z axis. When the PZT and the probe move down, the spiral metal plate will deform under such constraints. However, the displacement of the probe is less than the PZT because of the reaction of the tissue. Under the same force, the softer the tissue, the greater the deformation of the tissue, which can be seen in Fig. 4. Four samples with elasticity modulus of 100kPa, 200kPa, 300kPa and 400kPa are used to simulate the deformation. The indentation of tissue gets shallower and shallower as predicted above. The actual contact is well simulated by result of the static structural analysis, which is the basis of the following coupling analysis and the comparison of the experimental results.Modal analysis To evaluate the natural frequencies and the mode shapes of the tactile sensor, a modal analysis is performed, the results are shown in Table 2 and Fig. 5.
The only constraint in the modal analysis is the fixation of the side face of the PZT, as shown in Fig. 5 (a). The first natural frequency is 230.02Hz, far less than the natural frequency of the PZT. At such frequency, the impact brought by the effective mass and the low-frequency perturbation can be reduced, which means the first natural frequency is suitable for measuring the hardness of biological tissue. The other three frequencies are obviously larger than 230.02Hz, and the deformation of the probe is offset from
the vertical direction (Z axis), which cannot be picked. The result of the modal analysis can be used as a reference for the harmonic response analysis. 3.2 Harmonic response analysis The harmonic response analysis, taking into account the result of the static structural analysis, is carried out, in order to obtain the electrical impedance of the PZT when the tactile sensor touches the tissue. The material of the ceramic is PZT-5A, whose physical parameters can be find in appendix. Four silicone specimens with different hardness values are tested, as shown in Table 3. An alternating voltage with varying frequency (200 to 300Hz) is applied on the polarization surface of the PZT. The frequency range is chosen by referring to the results of the modal analysis, where the tactile sensor will resonate near the first natural frequency. The electrical impedance signals corresponding to the four silicone samples can be obtained by solving the harmonic response analysis, as shown in Fig. 6. Abrupt change of the amplitude of each signal occurs at a particular frequency, which is exactly the resonant frequency. The resonant frequencies corresponding to four samples of different hardness are 205Hz, 224Hz, 233Hz and 246Hz, respectively. The significant differences between the four frequencies show the sensor’s ability of measuring hardness, which verifies the feasibility of the tactile sensor. Then, 14 silicone samples of different hardness values are analyzed in order to calibrate the sensor, Table 4 and Fig. 7 shows the hardness of the 14 samples and the calibration curve respectively. The curve is fitted in MATLAB, the formula of fitting curve is: 𝑓𝑠 (𝑥) = −3435𝑥 −0.8432 + 254.3
(4)
where 𝑥 stands for the elasticity modulus in kPa, and 𝑓𝑠 (𝑥) stands for the simulating resonant frequency in Hz. The coefficient of determination of the fitted curve is 99.89%. The first half of the curve is steeper, which means the tactile sensor is sensitive to the tissue whose elasticity modulus is less than 1.5MPa. This range of elasticity modulus is consistent with that of most biological soft tissues, which means the proposed tactile sensor is suitable for the hardness measurement of biological tissue. 4. Experiments and results For further verification of the sensor’s feasibility, a prototype of the sensor is fabricated and an experimental system is set up, as shown in Fig. 8 and Fig. 9. A data acquisition card (National Instruments, NI USB-7856R) connecting to a computer is used as a signal generator and a data collector, controlled by a data acquisition and analysis software based on LabVIEW. The tactile sensor is fixed to the linear stage by a fixture. An oscilloscope is used to show the feedback signal more intuitively and the electronic balance is used to control the contact force. The measurement process can be divided into the following steps: First, a contact force of 30mN is generated between the sensor and experimental samples. Then, an alternating voltage signal is applied to the PZT and the current flowing through is used as the feedback signal, measured by an amplifying circuit and acquired by the data acquisition card. After that, the electrical impedance of the PZT can be calculated from the voltage and current signal, which is related to the mechanical impedance of the whole sensing system. Finally, the resonant frequency can be extracted from the electrical impedance, which reflects the information of hardness. 4.1 Hardness measurement The performance of the tactile sensor is evaluated on five silicone samples, whose hardness changes according to concentration. Fig. 10 and Table 5 show the five silicone samples and the hardness of them respectively.
In order to compare with the simulation results, the sensor is fixed to the linear stage, similar to the constraint in simulation. The natural frequency of the sensor itself is measured first, which is 232.3Hz, very close to the result of the modal analysis, which is 230.02Hz. Then, the contact force between the sensor and the sample is controlled by the displacement of the linear stage. As 30mN is set in the static structural analysis, the displacement is controlled to make sure the actual contact force is 30mN, which can be shown in the electronic balance. Each sample is measured 20 times and the results are shown in Fig. 11, comparing with the simulation fitting curve. The experimental results show the sensor performs well in hardness measurement. The data are in good agreement with the simulation curve. A small deviation in higher modulus range is observed between the FEM simulation result and the experimental result. This is due to the differences between the actual condition and the simulation setting, such as the material characteristics, the manufacturing errors and meshing accuracy. To meet the actual measurement results, the calibration curve of the sensor is corrected by the experimental data, as shown in the following formula: 𝑓𝑒 (𝑥) = −2305𝑥 −0.7243 + 266.6
(5)
where 𝑥 stands for the elasticity modulus in kPa, and 𝑓𝑒 (𝑥) stands for the experimental resonant frequency in Hz. The coefficient of determination of the experimental fitting curve is 99.91%. In this paper, the PZT used in the prototype of the sensor is a general product, whose size is small and the thickness is large, leading to the large impedance of PZT itself and the small amplitude of the feedback current signal. So the feedback signal can be easily affected by the noise signal of the sensing system, which reduces the reliable measuring range of the sensor. As to the sensor presented in this paper, the actual reliable measuring range of elasticity modulus is smaller than 500kPa, which is narrower than the simulating measurement range because of the impact brought by the noise. However, the accuracy of the measurement can be improved and the reliable measuring range of the sensor can be expanded by increasing the signal-to-noise ratio. In the next step of this research, the PZT will be made by customization to enlarge its size and reduce the thickness to reduce its impedance. A filter and a signal amplifier will be added to the sensing system. 4.2 Lump detection After verifying the sensor’s ability of measuring hardness, another silicone sample with a lump inside is made to test whether the lump can be detected by the sensor. As shown in Fig. 12(a), the silicone sample is 30mm in diameter and 15mm thick, a coin is used as the lump locating about 1mm below the surface of the silicone sample.
In order to compare with the experimental fitting curve, a force of 30mN is applied to the sample and several points of each area are tested. The average of the results is marked in the Fig. 12(b), which is 206.3Hz for the area without lump and 223.5Hz for the area with lump. Each area is marked with red for different depths corresponding to the frequency values, the higher the value, the deeper the color. The elasticity modulus of each area can be calculate referring to Eq. 5, which is 151.8kPa and 246.3kPa respectively. The results show the sensor’s ability of lump detection and the deeper lump can be detected if the contact force becomes larger. 5. Conclusion A new tactile sensor for measuring the hardness of soft tissue is presented. The proposed tactile sensor consists of a piezoelectric ceramic plate, a spiral metal plate and a probe. Only one piece of PZT is used as both actuator and sensing element, which reduces
the amount of sensor’s components. A simple spiral metal plate is designed to reduce the resonant frequency of the sensor for restricting the impact brought by the effective mass of the tissue. The structure of the sensor is very simple, which can be easily miniaturized. The concept structure of the sensor is proposed and working principle is explained. The relationship between resonant frequency and hardness of tissue is discussed by using Kelvin viscoelastic model. FEM analysis including static structural analysis, modal analysis and harmonic response analysis are carried out to verify the feasibility of the sensor and compare with the experimental results. The contact model of the sensor and tissue is obtained by static structural, which is also the basic of the harmonic response analysis. The first four resonant frequencies are solved by the modal analysis, which is the reference for selecting frequency sweep range. The sensor’s ability of measuring hardness is verified and the simulation fitting curve is obtained by harmonic response analysis. A prototype of the sensor is fabricated and the experimental system is established, five silicone samples are used to test the sensor. The feasibility of the sensor is further verified by the experimental results, which has a good agreement with the simulation results. The simulation fitting curve is also corrected by the experimental results. Furthermore, the ability of lump detection is demonstrated by a silicone sample with a lump inside. The application and evolution of this tactile sensor will be helpful for the development of MIS. Acknowledgment This work is supported by the National Natural Science Foundation of China (No.51575256) and the Starting Fund of Nanjing University of Aeronautics and Astronautics (No. 90YAH16062). REFERENCES [1]
Bicchi A, Canepa G, De Rossi D, et al. A sensor-based minimally invasive surgery tool for detecting tissue elastic properties[C]// IEEE International Conference on Robotics and Automation, 1996. Proceedings. IEEE, 1996:884-888 vol.1.
[2]
Dargahi J, Parameswaran M, Payandeh S. A micromachined piezoelectric tactile sensor for use in endoscopic graspers.[J]. Microelectromechanical Systems Journal of, 2000, 9(3):329-335.
[3]
Araghi M H, Salisbury S P. Improved Evaluation of Dynamic Mechanical Properties of Soft Materials With Applications to Minimally Invasive Surgery[J]. IEEE/ASME Transactions on Mechatronics, 2013, 18(3):973-980.
[4]
Eltaib M E H, Hewit J R. Tactile sensing technology for minimal access surgery––a review[J]. Mechatronics, 2003, 13(10):1163-1177.
[5]
Kalanovic D, Ottensmeyer M P, Gross J, et al. Independent testing of soft tissue visco-elasticity using indentation and rotary shear deformations.[J]. Studies in Health Technology & Informatics, 2003, 94:137-143.
[6]
Liu H, Noonan D P, Challacombe B J, et al. Rolling Mechanical Imaging for Tissue Abnormality Localization During Minimally Invasive Surgery[J]. IEEE Transactions on Biomedical Engineering, 2010, 57(2):404-414.
[7]
Liu H, Li J, Song X, et al. Rolling Indentation Probe for Tissue Abnormality Identification During Minimally Invasive Surgery[J]. IEEE Transactions on Robotics, 2011, 27(3):450-460.
[8]
Kalantari M, Shen J J, Dargahi J, et al. Localization of annulus with a tactile sensor[C]// IEEE, Northeast Bioengineering Conference. IEEE, 2011:1-2.
[9]
Li C G, Shen J J. Design and analysis of a tactile sensor used in minimally invasive surgery[C]// Ieee/asme International Conference on Advanced Intelligent Mechatronics. IEEE, 2013:1454-1457.
[10] Kanda T, Morita T, Kurosawa M K, et al. A flat type touch probe sensor using PZT thin film vibrator[J]. Sensors & Actuators A Physical, 2000, 83(1–3):6775. [11] Barthod C, Teisseyre Y, Géhin C, et al. Resonant force sensor using a PLL electronic[J]. Sensors & Actuators A Physical, 2003, 104(2):143-150. [12] Beekmans S V, Iannuzzi D. Characterizing tissue stiffness at the tip of a rigid needle using an opto-mechanical force sensor[J]. Biomedical Microdevices, 2016, 18(1):1-8.
[13] Valtorta D, Mazza E. Dynamic Measurements of Soft Tissue Viscoelastic Properties with a Torsional Resonator Device[M]// Medical Image Computing and Computer-Assisted Intervention – MICCAI 2004. Springer Berlin Heidelberg, 2004:481-490. [14] Ju F, Ling S F. A micro whisker transducer with sensorless mechanical impedance detection capability for fluid and tactile sensing in space-limited applications[J]. Sensors & Actuators A Physical, 2015, 234:104-112. [15] Omata S, Terunuma Y. New tactile sensor like the human hand and its applications [J]. Sensors & Actuators A Physical, 1992, 35(1):9-15. [16] Murayama Y, Haruta M, Hatakeyama Y, et al. Development of a new instrument for examination of stiffness in the breast using haptic sensor technology[J]. Sensors & Actuators A Physical, 2008, 143(2):430-438. [17] Omata S, Murayama Y, Constantinou C E. Real time robotic tactile sensor system for the determination of the physical properties of biomaterials[J]. Sensors & Actuators A Physical, 2004, 112(2):278-285. [18] Jalkanen V, Lindahl O A. Hand-held resonance sensor instrument for soft tissue stiffness measurements - a first study on biological tissue in vitro[M]// XII Mediterranean Conference on Medical and Biological Engineering and Computing 2010. Springer Berlin Heidelberg, 2010:463-466. [19] Åstrand A P, Jalkanen V, Andersson B M, et al. A Flexible Sensor System Using Resonance Technology for Soft Tissue Stiffness Measurements –Evaluation on Silicone[M]// 15th Nordic-Baltic Conference on Biomedical Engineering and Medical Physics (NBC 2011). Springer Berlin Heidelberg, 2011:21-24. [20] Jalkanen V, Andersson B M, Lindahl O A. Stiffness of a small tissue phantom measured by a tactile resonance sensor[M]// XII Mediterranean Conference on Medical and Biological Engineering and Computing 2010. Springer Berlin Heidelberg, 2010:395-398. [21] Åstrand A P, Andersson B M, Jalkanen V, et al. Initial Measurements on Whole Human Prostate ex vivo with a Tactile Resonance Sensor in Order to Detect Prostate Cancer[J]. Ifmbe Proceedings, 2015, 48:120-123. [22] Åstrand A P, Jalkanen V, Andersson B M, et al. Stiffness measurements on spherical surfaces of prostate models, using a resonance sensor[M]// World Congress on Medical Physics and Biomedical Engineering May 26-31, 2012, Beijing, China. Springer Berlin Heidelberg, 2013:1401-1404. [23] Dhar P R, Zu J W. Design of a resonator device for in vivo measurement of regional tissue viscoelasticity[J]. Sensors & Actuators A Physical, 2007, 133(1):4554.
Biographies Lei Zhang received the Bachelor’s degree from Nanjing University of Aeronautics and Astronautics, Nanjing, China in 2015. He is currently a Master’s degree candidate at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include tactile sensors and surgical robots. Feng Ju received the PhD degree from Nanyang Technological University, Singapore. He is currently an Associate Professor in the School of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics. His research interests include piezoelectric sensors and actuators, robotic tactile sensing and motion control for robotics. Yanfei Cao received the Bachelor’s degree from Nanjing University of Aeronautics and Astronautics, Nanjing, China in 2016. He is currently a Master’s degree candidate at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include surgical robots and cable-driven robots. Yaoyao Wang received the Bachelor’s degree from Southeast University, Nanjing, China in 2011 and the PhD degree from Zhejiang University, Hangzhou, China in 2016.He is currently a Lecturer in the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include robust control and adaptive control of underwater vehicle-manipulator system, design and control of cable-driven manipulators. Bai Chen received the Bachelor’s and PhD degrees from Zhejiang University, Hangzhou, China, in 2000 and 2005, respectively. He is currently a Professor and PhD candidate supervisor at the College of Mechanical and Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China. His current research interests include design and control of surgical robots, cable-driven robots and sperm-like swimming micro robots.
Spiral metal plate
PZT
Constraint surface
Probe
Fig. 1. Concept structure of the tactile sensor
Piezoelectric Actuator
k1
F
c1 Indenter Head
mtip mm
x km
cm
Tissue
Fig. 2. Connect model of the tactile and tissue
PZT Metal plate
Probe
Fig. 3.
Deformation of the tactile sensor under harmonic excitation
(a)
(b)
(c)
(d) Fig. 4. Results of the static structural analysis: (a) sample with elasticity modulus of 100kPa, (b) sample with elasticity modulus of 200kPa, (c) sample with elasticity
modulus of 300kPa, (d) sample with elasticity modulus of 400kPa
(a)
(b)
(c)
(d) Fig. 5. First four resonant modes: (a) the first mode, (b) the second mode, (c) the third mode, (d) the forth mode
Fig. 6.
Electrical impedance signals corresponding to the four silicone samples
Fig. 7.
The fitting curve fitted by 14 simulating silicone samples
PZT
Probe
Spiral metal plate Fig. 8.
Signal wire 1cm
The prototype of the sensor comparing with a scale
Fig. 9.
The experimental system
Sample# A
Sample# B
Sample# D
Sample# C
Sample# E
Fig. 10. Five silicone samples with different hardness values
Fig. 11. Experimental results comparing with simulation fitting curve
206.3Hz
Points detected
Lump
Silicone sample 223.5Hz
(a)
(b) Fig. 12. Lump detection: (a) the silicone sample with lump inside, (b)
the detecting result
TABLE. 1.
TABLE. 2.
TABLE. 3.
Material of each component
Component
Material
PZT
PZT-5A
Spiral metal plate
Stainless steel
Probe
Stainless steel
The first four resonant frequencies
Mode
Frequency (Hz)
1
230.02
2
326.31
3
390.96
4
1331.9
Four silicone specimens
Code
Shore hardness (HA)
Elasticity modulus (kPa)
A
0
157.5
B
5
279
C
10
414
D
30
1146
TABLE. 4.
Elasticity modulus of 14 silicone samples
Code
TABLE. 5.
Elasticity
modulus
(kPa)
Code
Elasticity
modulus
(kPa)
1
100
8
800
2
200
9
900
3
300
10
1000
4
400
11
1500
5
500
12
2000
6
600
13
2500
7
700
14
3000
Five silicone samples for experimental study
Code
Shore hardness (HA)
Elasticity modulus (kPa)
A
0
157.5
B
5
279
C
10
414
D
30
1146
E
50
2465
APPENDIX 1. The expressions of the components of tissue’s impedance When the sensor presented in this paper is excited to vibrate at its natural frequency and pressed to the tissue, as shown in the Fig. 13, the impedance of the tissue can be expressed as:
𝑍𝑚 = 𝑐𝑚 + 𝑗𝜔𝑚𝑚 +
𝑘𝑚 𝑗𝜔
where 𝑐𝑚 , 𝑚𝑚 and 𝑘𝑚 are equivalent damping, equivalent mass and equivalent stiffness respectively. They can be expressed as:
𝑐𝑚 =
𝐶0 𝜌𝑚 𝑣𝑚 𝑆𝑐 1 − 𝜐𝑚 3
0.1𝜌𝑚 𝑆𝑐2 𝑚𝑚 = 1 − 𝜐𝑚 {
𝑘𝑚 =
2𝐸𝑚 √𝑆𝑐 2) √𝜋(1 − 𝜐𝑚
where 𝐶0 is a function related to the Poisson ratio 𝜐𝑚 , which can be seen as a constant. 𝜌𝑚 is the density of the tissue, 𝑆𝑐 =π𝑎02 is the contact area, 𝑎0 is the radius of the contact area, 𝐸𝑚 is the elasticity modulus of the tissue and 𝑣𝑚 = √ velocity of the wave spread in the tissue.
Z1
Sensor
T issue
Zm
2
a
0
E ffec tiv e mass
Fig. 13 Contact modal of the measurement
2. The material parameters of PZT-5A The material of the PZT set in FEM analyses is PZT-5A, whose physical parameters are: 𝑑=[
0 0 −5.35
0 0 −5.35
0 0 15.78
0 12.26 0
12.26 0 0
0 0] × 10−10 𝐶/𝑁 0
𝐸𝑚 2(1+𝜐𝑚 )𝜌𝑚
is the
12.04 7.52 7.51 𝐸 𝑐 = 0 0 [ 0
7.52 12.04 7.51 0 0 0
7.51 7.51 11.09 0 0 0
8.13 𝜀𝑆 = [ 0 0
0 8.13 0
0 0 0 2.26 0 0
0 0 0 0 2.11 0
0 0 0 × 1010 𝑁/𝑚2 0 0 2.11]
0 0 ] × 10−9 𝐹/𝑚 7.31
where 𝑑, 𝑐 𝐸 , and 𝜀 𝑆 are the piezoelectric matrix, the stiffness matrix and the dielectric matrix, respectively.