ADAPTIVE NEURAL NETWORK CONTROL OF A HUMAN SWING LEG AS A DOUBLE-PENDULUM CONSIDERING SELF-IMPACT JOINT CONSTRAINT

Abstract

For human walking, the swing leg is usually modeled as a double pendulum. Considering a joint self-impact constraint at the knee joint of the double pendulum model is the main difference in this study. The primary objective of this research is to propose a nonlinear Adaptive Neural Network (ANN) for this system. By using Gaussian RBF networks, asymptotically stable tracking is attained. We will use the available data of normal human walking for the desired trajectories of the hip and knee joints. By simulation of the system, we perceive that the swing leg tracks the normal human gait with a negligible and tolerable error.