A self-powered magnetoelectric tactile sensor for material recognition

Abstract

Due to the complex working environments, robots need to sense and react to their various interactive surroundings with the help of tactile perception. Material recognition, one of the basic applications of tactile perception, can be achieved by further interpreting and representing the tactile data collected by tactile sensors. Suitable control strategies can be implemented to ensure stable and efficient operation of robots if the contacting materials are accurately recognized. Therefore, in this paper, a soft, deformable, and self-powered tactile sensor based on magnetoelectric materials is developed. Structural design, working principles, and shape optimization of this magnetoelectric tactile sensor are first described. Fabrication and assembly of the magnetoelectric tactile sensor are then introduced. The material recognition experiments including the voltage ratio ranges study, test 1 (velocity change), and test 2 (size and shape change) are performed to validate the effectiveness of the proposed magnetoelectric tactile sensor. The experimental results show that the recognition accuracy is high (around 90% for soft materials and 100% for hard materials) regardless of sizes and shapes of the contacting materials.