Muscle Fatigue Detection Method for Exoskeleton Robot Operation Based on Electrical Impedance Imaging Enhancement *

Due to the lack of human–robot coordination, long-term improper operation of exoskeleton robots leads to muscle fatigue and injury. However, current muscle fatigue detection methods suffer from noise interference and are unable to achieve visualization. To dynamically monitor muscle fatigue during exoskeleton robot operations, a novel electrical impedance imaging enhancement (EI-IE) method is proposed. Firstly, the TK-Noser regularization algorithm is used to generate a vague image. Secondly, the DBSCAN clustering algorithm is employed to divide the TK-Noser image into n groups and assign different weights to each group. Finally, two weighted TK-Noser images are simultaneously fed into an intelligent optimization algorithm for iterative optimization. Simulation results demonstrate that the EI-IE results achieve an average Intraclass Correlation Coefficient (ICC) / Root Mean Squared Error (RMSE) of 0.823/0.254, which is 16.8 %/15.3 %, 3.1 %/7.9 % better than Tikhonov and TK-Noser, respectively. The experimental platform was established to verify the imaging capability of the EI-IE algorithm, with an average ICC/RMSE of 0.743/0.266, which is 19.7 %/15.3 % and 2.5 %/11.9 % better than Tikhonov and TK-Noser, respectively.