Development and Analysis of a Bio-Inspired Tactile Sensor

Kurzfassung

Recent advances in robotics and mechatronics enable the design of robotic devices with an increasing market potential, but the introduction to man’s everyday environment is still a long way to go. Following this path, currently an anthropomorphic Hand-Arm system is developed at DLR. Its desired small size and the mechatronic integration result in little designed space, preventing the application of “classical” sensor systems such as 6DoF load cells. Following the anthropomorphic design approach pursued in the design of the Hand-Arm system, a bio-inspired sensor setup is desired. Therefore this study aims at the development of a suitable sensor setup. As a result of the literature research on dextrous manipulation, robotic handhardware and tactile sensors, piezoresistivity has been chosen as transduction principle. For the derivation of the bio-inspired sensor setup the general requirements for valuable tactile sensors and the special requirements resulting form the anthropomorphic design approach are assessed. Hence a consistency matrix is applied to evaluate the requirements and to identify possible conflicts of goals between the requirements. To solve the conflicts of goals, the following bionic method is applied: First the the technical challenge is abstracted, secondly the equivalent functionality in human tactile perception is analyzed and the functional principle, i.e., nature’s strategy is abstracted. Where applicable, the functional principle is propagated to a technical system. Following possible technical approaches are derived. Finally, the proposed approaches are assessed using a rating matrix. Thus the the most promising approaches are identified and combined to a bioinspired sensor setup. The proposed bio-inspired sensor setup consists of a “dermal” and a superficial sensor which are arranged in different layers of an artificial skin. While the “dermal” sensor is based on metal readout wires cast into a compliant piezoresisitive material, the superficial sensor is designed as a strain sensitive layer covering the “dermal” sensor. For the investigation of the proposed sensor principles a modular testbed has been designed and realized within this study. In its current version the testbed allows for automated testing of a testpatch under a pre-set force or indentation depth. For the evaluation of different sensor materials, testpatches have been manufactured and examined on the testbed and with a tensile testing machine. Based on the results of these tests, a particularly suitable piezoresistive silicone has been selected for the further development of the tactile sensor. During the development of the “dermal” sensor the conducted tests revealed an optimization challenge concerning good piezoresistive behavior, high compliance, and good adhesion to the cast-in metal wires. As the means of optimization are in conflict the proposed principle for the “dermal” sensor could not be proven. Although the combination with metal wires is not applicable, the evaluated silicone shows very promising properties, being well-suited for the superficial sensor. Based on the results of the experiments for the “dermal” sensor, a new transduction principle for the superficial strain sensor was developed and patented. Applying the afore investigated materials, a prototype of the superficial sensor could be realized. In a first set of tests the sensor prototype showed an exponential correlation between the applied force and the measurable resistance, and could easily be linearized with a logarithmic amplifier. As the sensor principle for the superficial strain sensor could be proven, continuative studies will deal with the redesign of the “dermal” sensor and the propagation of the bio-inspired sensor setup to the anthropomorphic fingertip.