Wearable Flexible Optical Fiber Sensors for Human Motion Intention Recognition

Dong Liu,Yuxuan Zhang,Nian Peng,Haojie Liu,Wei Meng,Quan Liu

DOI: https://doi.org/10.1109/M2VIP58386.2023.10413443

2023-01-01

Abstract:In daily life, it is inevitable to encounter situations such as joint and muscle injuries. In order to maximize the recovery of the affected area, rehabilitation training is indispensable. Currently, the acquisition of human biometric signals is essential in the process of rehabilitation training. Commonly used devices for collecting human biological signals exhibit certain limitations, such as difficult to resist electromagnetic interference and lack of privacy. In response to the aforementioned issue, we propose a fiber Bragg grating deformation and angle sensor, which has the characteristics of high flexibility, high sensitivity, low hysteresis and easy manufacturing. We derived the sensing principle of the sensor, conducted mechanical performance testing, and performed consistency analysis with electromyographic signals. a convolutional neural network was employed to discern human movement intentions from the acquired data using FBG-DAS. The results indicate that the sensor exhibits excellent mechanical characteristics, high consistency with electromyographic signals, and achieves a classification accuracy of 93% in recognizing human movement intentions, such as different walking speeds, standing up, and sitting down, during the experiment.