Model-based Control with Interaction Predicting for Human-coupled Lower Exoskeleton Systems

Abstract

Sensitivity Amplification Control (SAC) algorithm was first proposed in the augmentation application of the Berkeley Lower Extremity Exoskeleton (BLEEX). Since the SAC algorithm can greatly reduce the complexity of exoskeleton system, it is widely used in human augmentation applications. Nevertheless, the performance of the SAC algorithm depends on the accuracy of dynamic model parameters. In this paper, we propose a novel Model-based control with Interaction Predicting (MIP) strategy to lower dependency on the accurate dynamic model of the exoskeleton. The MIP consists of an interaction predictor and a model-based controller. The interaction predictor can predict motion trajectories of the pilot and substitute for the pilot to drive the exoskeleton through an impedance model. In proposed strategy, the model-based controller not only amplify the forces initiated by the interaction predictor, but more importantly the forces imposed by the pilot to correct the errors between the predictive motion trajectory and the intended motion trajectory of the pilot. Illustrative simulations and experimental results are presented to demonstrate the efficiency of the proposed strategy. Additionally, the comparisons with traditional model-based control algorithm are also presented to demonstrate the efficiency and superiority of the proposed control strategy for lowering dependency on dynamic models.