Control of Walking in Unilateral Drop Foot Using Artificial Neural Networks and Hybrid LMA-PSO Training Algorithm (original) (raw)

Biosystems & Biorobotics, 2013

Abstract

ABSTRACT The goal of this paper is to control the FES intensity for the disabled tibialis anterior (TA) muscle in patients with unilateral drop foot, using the existing coordination patterns between the activities of ipsilateral ankle dorsiflexor muscles and contralateral ankle plantarflexor muscles during normal gait. A nonlinear auto- regressive neural network (NARX) and an adaptive-network-based fuzzy inference system (ANFIS) are trained to forecast the TA muscle activity of one foot using gastrocnemius muscle activity of contralateral foot. A two level efficient hybrid training algorithm is employed which consists of a gradient based level (local search) and an evolutionary level (global search). The predicted TA activation then can be used to control the TA muscle FES intensity in real time, using a linear mapping. Eleven healthy volunteers participated in the experiments. Quantitative evaluations show the promising performance of developed neural controllers.

hamid reza kobravi hasn't uploaded this paper.

Let hamid reza know you want this paper to be uploaded.

Ask for this paper to be uploaded.