Multifunctional\nSelf-Powered Sensors Integrated on\na Robot Hand for Detecting Temperature–Pressure Stimuli and\nRecognizing Objects

Tactile\nsensing, especially pressure and temperature recognition,\nis crucial for both humans and robots in identifying objects. The\ngeneral solutions, which use piezoresistive, capacitive, and thermal\nresistance effects, are usually subject to single-mode sensing and\nan energy supply. Here, we propose a multimode self-powered sensor.\nThe sensor can respond to pressure and temperature stimuli using triboelectric\nand thermoelectric effects. Furthermore, we developed a sensing system\ncomprising sensors, a deep learning block, and a smart board. The\ndeep learning model can fuse features of triboelectric and thermoelectric\nsignals, enabling a high accuracy of 99.8% in recognizing ten objects.\nThis method may provide the future design of self-powered sensors\nfor object recognition in robotics.