• capacitive sensor
• multiplexed capacitance readout
• tactile sensor array

The objective of this Master Thesis is the development a Micro Electro Mechanical System (MEMS)-based capacitive tactile sensing system, intended to be integrated in the distal phalanx of a robotic finger. The used sensor-array consists of sixteen capacitive force sensors, two resistive temperature sensors and two reference capacitors. The prototype is capable of measuring the application of local forces and temperature changes. The sensors are silicon-based and are fabricated by means of the Bonded and Etched-Back Silicon-On-Insulator (BESOI) technology, and are the result of a collaboration between The BioRobotics Institute of Scuola Superiore Sant’Anna and the School of Mechanical Engineering of University of Birmingham, in the framework of the EU-funded (FP7-NMP) Nanobiotouch project. The sensing principle is capacitance-based because of the advantages in terms of higher spatial resolution, long term drift stability, increased sensitivity, lower temperature sensitivity and power consumption as compared to piezoresistive, strain gauge and piezoelectric sensing principles. The readout electronics is implemented by means of four capacitance-to-analog converters (Irvine Sensors, MS3110), with time-division multiplexing via low-capacitance CMOS quad-channel switches (Analog Devices, ADG1212) along each row of sensors of the array. The MS3110 output voltage is a function of the differential capacitance provided as an input to the converter, and of configuration parameters having effect on the conversion gain and offset. Such configuration parameters of the MS3110 capacitance-to-analog converter are programmed by means of the MS3110BDPC Evaluation/Programming Board.
During the Master Thesis, several experimental sessions were carried out in order to assess the subsystems of the tactile sensing system under study:
1. interfacing of readout electronics for the single-sensor;
2. interfacing of multiplexed electronics for the sensor-array;
3. probing of individual capacitive sensor unit.
Within the first set of experimental activities, a single-sensor readout electronics was tested in order to evaluate the functionality of the MS3110 capacitance-to-analog converter: discrete capacitances, in the range between 0.25pF and 10pF, were tested as a function of the values of the converter parameters setting the gain and the offset. The values of the output voltage were recorded and elaborated as a function of the difference between the sensing capacitances and of the used configuration parameters, to verify the nominal characteristics of the capacitance-to-analog converter.
Within the second set of experimental sessions, the interfacing of the multiplexed readout electronics for the sensor-array was tested in order to evaluate its functionality in measuring the capacitance of the 16 channels of the array, under a time-division policy. The signal to the capacitance-to-analog converter is switched between discrete capacitances using the quad-channel switches (ADG1212). Within this experimental protocol, the performance of one of the four measurement circuits was evaluated while operating a time-division multiplexing via the switch along each row of the measurement circuit. Discrete capacitances, in the range between 0.25pF and 10pF, were tested as a function of the values of the converter parameters. The experimental results show a general consistency between the nominal characteristics of the capacitance-to-analog converter and the experimental data.
Within the third set of experimental activities, the mechanotransduction properties of the sensor were evaluated by direct probing of individual capacitive sensor units, i.e. by characterizing the response to indentation cycles. A loading system allowed precise positioning of a loading probe endowed with a spherical head: the contact between the probe and the sensor was first adjusted with three manual orthogonal micrometric translation stages (M-105.10, PI), and then precisely controlled along the indentation direction by means of a servo-controlled positioning system (M-111.1, PI); two camera permitted the proper positioning of the probe over the sensor. To measure the applied force, a six-component load cell was positioned at the interface between the probe and the translational stage. The values of output voltages were recorded and elaborated as a function of the applied load, to evaluate the characteristics of the capacitive tactile sensor. The results of this experimental protocol show, coherently with the expectations, that the sensor and the readout electronics responded to an increase in probe indentation depth with an increase in measured voltage change.