Control of Assistive Lower Limb Exoskeleton with Joint Torque Estimation with Momentum Observer

Abstract

Wearing a lower limb exoskeleton is a promising solution for assisting patients or senior people with mobility problems. It is vital to detect human intention because it can help to effectively and safely assist a wearer’s motion. However, the techniques currently available for detecting human intention using surface electromyographic (sEMG) and estimating the human gait phase are unreliable in certain circumstances such as irregular gait patterns, climbing stairs, sudden stops, and changes in direction. This study proposes, a gait assistance control strategy for overcoming the afore mentioned limitations. The technique comprises three steps. In the first step, a combined model of the exoskeleton and its wearer was designed as a five-link model. In the second step, to detect human intention in a limited fashion, the joint input torques of the combined model required for walking were estimated by using a momentum observer based on the dynamics of the combined model. In this step, only the angular displacement and joint velocity measurements were required to estimate the joint input torques. The final step involved designing an exoskeleton controller for generating assistance torques by applying the fictitious gain (FG) method. This process allowed the exoskeleton controller to amplify the human joint torques. A computer simulation was performed to validate the effectiveness of the proposed control strategy. In terms of reducing human torque, the results obtained were satisfactory.