Design of Assistive Torque for a Lower Limb Exoskeleton Based on Motion Prediction

Abstract

This paper presents a method to generate assistive torque to a lower limb exoskeleton using a predictive approach. A control scheme is proposed to predict the suitable force of next movement of legs using gait information. By using human walking gait cycle and Center of Pressure (CoP), it is expected that the exoskeleton can provide an assistive force to people and help them to walk normally. In order to predict the next movement for assistive torque given by exoskeleton to the human, a predictive Artificial Neural Network (pANN) is developed. A prototype lower limb exoskeleton has been designed and constructed for experimental study. The controller is implemented by using LabVIEW programming. The experimental results verified that the proposed controller can provide proper assistive torques to hip and knee joints for both legs while wearing the exoskeleton.