Multichannel fused EMG based biofeedback system with virtual reality for gait rehabilitation

Abstract

Aim of this paper is to develop an EMG based biofeedback system using a virtual reality platform which will help in gait rehabilitation. A low power multichannel EMG acquisition unit has been developed to acquire EMG of six different muscles of the lower limb. EMG from different channels are fused using Bayesian fusion technique and spurious data has been discarded. From the fused EMG data, we calculate different gait parameters like stride time, gait phase etc. Joint trajectory during a gait cycle is obtained, digitized and combined with the gait parameters acquired from EMG. Together they are fed to a VR human model. Just like a person walks, the same EMG and trajectory data being fed to the model, it walks too mimicking the gait of the user, with the same speed, thus providing biofeedback to the user.The system has massive application in gait rehabilitation for poststroke patients, people suffering from cerebral palsy and other neuro muscular gait defects, amputees etc.

High SNR EMG Acquisition System for Biofeedback Applications

“…If the application demands superior SNR performance, then high end electrodes are the optimum solution. But this is not mandatory for many applications like those discussed in literatures [11], [12], [13], [14], [15]. Electro-conductive gel should be applied before placing the electrodes or wire to get better performance.…”

Generating gait pattern of myoelectric active ankle prosthesis

“…(6). More details regarding fusion technique can be found in [9]. Activeness in the joint demands some form of actuation to provide the required power.…”

“…Two parameters are commonly used to measure the amplitude, the root-mean-square (RMS) value and the mean absolute (MA) value. We have calculated moving window RMS with adaptive mean selection algorithm [9]. Moving window RMS detection technique is applied to compute RMS for each successive incoming EMG data sample.…”

Surface Electromyography Based Finger Flexion Recognition

“…It was done by extracting features consisting integrated EMG, waveform length, zero crossing and Willison amplitude. Biswas [14] proposed a method for imitating gaits. EMG from different channels taken from the lower limb are fused using Bayesian fusion technique.…”