Analysis of gait-pattern adaptation algorithms applied in an exoskeleton for lower limbs

This paper presents the analysis of three gait pattern adaptation algorithms applied to trajectory control of an exoskeleton for lower limbs. The proposed exoskeleton is based on a commercially available orthosis. The main dynamic characteristics of the orthosis-patient system are considered in MatLab and ADAMS simulations. Also, three gait-pattern adaptation algorithms are implemented considering the proposed model. These algorithms lead to a reduction of the estimated interaction forces between the human and the robot, allowing patients to change the gait-pattern according to their degree of voluntary locomotor capability. The first one is based on the inverse dynamics and minimizes a functional computed from interaction forces and orthosis-patient dynamics. In the second, based on the direct dynamics, the human-robot interaction forces are translated into a desired change in trajectory accelerations. The third one minimizes the interaction forces by controlling the impedance between the orthosis and the patient. Experimental joint trajectories obtained from the orthosis are considered as initial trajectories for the gait-pattern algorithms.