This manuscript presents the modeling, simulation, and control of a planar cable-driven lower-limb rehabilitation robot. A computed-torque controller is used as the model-based baseline, and a residual DDPG reinforcement learning controller is added to improve robustness under randomly generated disturbances and parametric uncertainties. Simulation results show an approximately 40 percent reduction in tracking error compared with the baseline controller.
@article{fakouri2026rehabrobot,
title = {Control and Simulation of a Cable-Driven Lower-Limb Rehabilitation Robot Using Residual Reinforcement Learning},
author = {Fakouri, Mohammad Hossein and Keymasi Khalaji, Ali},
journal = {Manuscript in preparation},
year = {2026}
}