Estimation of Desired Motion Intention and Compliance Control for Upper Limb Assist Exoskeleton

In this paper, we have addressed two issues for upper limb assist exoskeleton. 1) Estimation of Desired Motion Intention (DMI); 2) Robust compliance control. To estimate DMI, we have employed Extreme Learning Machine Algorithm. This algorithm is free from traditional Neural Network based problems such as local minima, selection of suitable parameters, slow convergence of adaptation law and over-fitting. These problems cause lot of problem in tuning the intelligent algorithm for the desired results. Furthermore, to track the estimated trajectory, we have developed model reference based adaptive impedance control algorithm. This control algorithm is based on stable poles of desired impedance model, forcing the over all system to act as per desired impedance model. It also considers robot and human model uncertainties. To highlight the effectiveness of the proposed control algorithm, we have compared it with simple impedance and target reference based impedance control algorithms. Experimental evaluation is carried on seven degree of freedom upper limb assist exoskeleton. Results describe the effectiveness of ELM algorithm for DMI estimation and robust tracking of the estimated trajectory by the proposed model reference adaptive impedance control law.