This paper proposes a novel stride management strategy for rehabilitation exoskeletons. This method incorporates a Central Pattern Generator (CPG) into an Assist-As-Needed (AAN) controller to induce an optimal stride length for the wearers and optimize the energy consumption. Most AAN controllers rely on a predefined and fixed trajectory to measure the required amount of assistance. However, for a stride management approach, the trajectory should be updated regularly according to the wearer’s performance to induce the optimal stride length eventually. The proposed stride management strategy deals with this challenge by integrating a CPG into the control loop. The CPG updates the desired trajectory right after each swing phase. The AAN controller uses a recently introduced Strength Index (SI) for continuous measurement of the wearer’s ability in tracking the desired trajectory. A virtual tunnel around the desired trajectory is defined, and the tunnel boundaries are adjusted according to the SI. Then, an adaptive impedance controller determines the assistive force according to the distance between the actual trajectory of the user and the tunnel boundaries. The performance of the proposed method is evaluated in OpenSim software, and reductions in the metabolic cost and muscles’ activity are observed.

Keywords