Disturbance Compensation Control of Lower Limb Rehabilitation Robot Based on RBF Neural Network and ESO

Lower limb rehabilitation robots have been increasingly applied in both clinical and home-based rehabilitation scenarios. However, during the rehabilitation training process, their control performance is susceptible to various factors such as friction, external interference, and model uncertainty. To tackle these difficulties, this paper proposes a disturbance compensation control algorithm based on RBF neural network and ESO. The control algorithm uses RBF neural network to learn online and approximate model uncertainty to achieve adaptive compensation for unknown dynamics. Simultaneously, the ESO estimates and compensates residual disturbances in real time, thereby improving the system’s robustness. Additionally, a hybrid control strategy that combines linear and nonlinear error feedback mechanisms is designed to improve the dynamic response of the system. The simulation results demonstrate that the proposed method achieves high-precision hip and knee joint tracking in terms of RMSE and MAE, ensuring stable and safe operation of lower-limb rehabilitation robots under complex environments.