GC-IGTG: A Rehabilitation Gait Trajectory Generation Algorithm for Lower Extremity Exoskeleton

Abstract

It is important to offer a natural and personalized rehabilitation gait trajectory, especially in the early stages of walking rehabilitation, for the patients with lower limb disability. Lower extremity exoskeleton has been proven to be efficient to provide highly repeatable and accurate rehabilitation exercise, but most existing exoskeletons’ gait trajectories won’t vary with the users. This paper proposes an algorithm, named as gait cell based individualized gait trajectory generation (GC-IGTG), for the purpose of offering a natural and personalized gait trajectory reference for the lower extremity exoskeleton based on the body parameters of the patients. The GC-IGTG is based on extreme learning machine and AutoEncoder, which makes it achieve fast training speed and suitable for small sample training conditions. The gait cell concept is proposed to improve the efficiency and safety of the algorithm. The experimental results indicate that the generated trajectories with GC-IGTG are almost identical to the original ones.