Averaged EMG Profile and Artificial Neural Network Based Walking Speed Estimation by GAIT Analysis

The data from an electromyogram (EMG) reveal the degree of muscle activation. In the fields of biomedicine, clinical practice, rehabilitation, and human-machine interface, EMG signals are becoming more and more valuable. Advanced technologies can be built for a variety of applications based on the features that are derived from the EMG signals. to create an artificial neural network model that uses EMG gait analysis to reliably estimate walking speeds. Using EMG signals, the Artificial Neural Network (ANN) may be used to accurately estimate walking speed. In this work, the neural network was trained using characteristics derived from averaged EMG profile data obtained from EMG signals from 14 muscles of 20 people at 5 different walking speeds. To determine the walking pace, a multilayer feed forward network model with a backpropagation method was created. The maximum, mean, and minimum accuracies of the developed model were determined to be 95.45%, 92.15%, and 80.0%, respectively. The estimated walking speeds using the suggested model were found to be satisfactory; however, the results might be improved to 100
^%
by removing additional features.