Fuzzy logic for walking patterns based on surface electromyography signals with different membership functions

Abstract

Classifying walking patterns is important in developing assistive robotic devices, especially for lower limb rehabilitation. Recently, Fuzzy Logic (FL) controllers have successfully been applied in grasping and control system for upper limb based on surface Electromyography (EMG) signals. Therefore, this paper evaluates the performance of FL with different membership functions in discriminating walking phases (e.g, stance and swing phases). The accuracy of two widely used membership functions (MF) like triangular and Gaussian is compared to identify their behavior for detecting the phases of walking. In this study, the MATLAB and Simulink toolboxes are used to examine the performance of each MF. Our findings show Gaussian MF gained better performance than the triangular MF with 90% of classification accuracy. Therefore, the Gaussian MF could be the best solution to classify the walking phases in this work.