Gesture Recognition Based on Nano-gold Flexible Sensor using Different Training Modes

摘要

Prosthetic control technology utilizing motion intentions decoded from surface electromyography (sEMG) signals is becoming more and more popular. Most of the traditional sEMG wet electrodes require the skin to be prepared by conductive gel, which could lead to skin allergy and patient discomfort. In this study, we proposed a new method of using nano-gold flexible sensors to measure muscle contraction in terms of changes in sensor impedance. The nano-gold flexible sensors were used to classify nine gestures in two training modes: the sequential training mode, in which the same motion was repeated six times, and the random training mode, in which the order of the motions was randomized. The results showed that the average gesture recognition rates of using nano-gold flexible sensors were above 90% for all the subjects participated in the experiments. There was no significant difference between the two training modes (94.54% for the sequential training mode and 94.16% for the random training mode), with a p-value of 0.7340. The study suggested that the nano-gold flexible sensors could be used as an alternative of the wet electrode for reliable gesture recognition.