Stretchable Self-Powered TENG Sensor Array for Human-Robot Interaction Based on Conductive Ionic Gels and LSTM Neural Network
Wentao Dong, Kaiqi Sheng, Bo Huang, Kun Xiong, Kun Liu, Xiao Cheng
Abstract
The flexible and stretchable self-powered sensors are widely applied to the Internet of Things (IoT), human motion monitoring, and human-robot interaction (HRI), and it reveals more and more attentions due to the flexibility and self-powered function. This article focuses on the stretchable triboelectric nanogenerator (TENG) sensor for human gesture monitoring and HRI application based on conductive ionic gels and machine learning. Stretchable self-powered TENG sensor is fabricated by solution method based on polyacrylamide (PAAM)/NaCl conductive hydrogel and silicone rubber. The self-powered TENG sensor array (TENG-SA) is applied to gesture detection at different parts (wrist, elbow, knee, and neck) of human body successfully without external power supply. Conductive PAAM/NaCl hydrogel could be stretched up to 307%, which satisfies the deformation of the human joints. Intelligent sensing system is designed and developed to control the motion of two-wheeled robot through TENG-SA and long and short time memory (LSTM) neural network via Bluetooth. Experiments have demonstrated that the neck motion signals are collected and recognized by intelligent TENG-SA system with LSTM neural network during the neck rotation process, which is applied to control the motion of two-wheeled robot successfully. HRI based on self-powered TENG sensor will promote the interaction among humans and machines/robots to improve the disabled for rehabilitation, medical monitoring, and human-robot cooperation.