Adaptive-oscillator-based control strategy for gait rehabilitation robots

Abstract

In the field of gait rehabilitation robotics, it is very important to provide synchronous assistance during walking. In this paper, a novel control strategy based on an adaptive oscillator is proposed. The adaptive oscillator estimates the stride percentage of gait based on the gait events information, which is detected in real time with a hidden Markov model (HMM). Synchronous reference trajectories for the robot are then generated with a look-up table. An impedance controller is implemented to provide assistance based on the synchronous reference. The proposed synchronization method is implemented in a portable knee-ankle-foot robot and tested in 15 healthy subjects. The experimental results demonstrate that the proposed control strategy is efficient in achieving human–robot synchronization and feasible for rehabilitation robotics application.