Phase-Based Repetitive Control for Rehabilitation Systems: A Case Study on Ankle Assistance

Active ankle foot orthoses (AAFOs) are widely used in gait rehabilitation to assist patients with impaired dorsiflexion and plantarflexion control. Due to the cyclic and quasi-periodic nature of human walking, repetitive disturbances and tracking errors frequently arise from actuator dynamics, transmission nonlinearities, and human device interaction forces. This article presents a phase based repetitive control (RC) framework specifically tailored for AAFO systems and integrated with a baseline impedance like controller to ensure safe and compliant assistance. By embedding an internal model of gait periodicity through phase synchronized learning, the proposed controller progressively reduces cycle to cycle tracking errors while remaining robust to cadence variations. A simulation case study demonstrates that the AAFO can learn a phase dependent assistive torque profile, resulting in improved tracking accuracy and smooth assistance under variable gait conditions. The results highlight the suitability of repetitive control as an interpretable and performance enhancing strategy for rehabilitation oriented biomechatronic devices.