Machine Learning-Assisted Intelligent Fibers for Remote Control

Abstract

Under the impact of Covid-19 virus, remote control is of value in non-contact systems. Glove-based wearable systems are promising for precise and low-cost hand gesture recognition. Yet, preparing stable intelligent fibers using facile techniques for reliable machine learning is still challenging. Here, we propose a stable intelligent fiber via layer-by-layer assemble for reliable machine learning, which exhibits a gauge factor of 4. The adoption of PVA and PU film can improve adherence of CNTs and stability of intelligent fiber during cyclic deformations, thus improving electrical performances and service time. Besides, integrating a flexible hybrid electronic system, we demonstrate remote control of robots using our fabricated glove and a shallow neural network.