Abstract—Control of Two-Dimensional (2D) rehabilitation robots is inherently challenged by nonlinearities, time-varying uncertainties, and modeling inaccuracies, which can significantly undermine compliance and tracking performance during human-robot interactions. To address these challenges, this paper presents a robust hybrid control strategy that integrates admittance control, adaptive Radial Basis Function (RBF) neural network compensation, and sliding mode control. Within this framework, admittance control is utilized to generate a compliant reference velocity based on the measured interaction force. The adaptive RBF neural network functions to estimate unmodeled nonlinear dynamics in real-time, operating without the need for prior system knowledge. Additionally, sliding mode control is employed to mitigate estimation errors and enhance system robustness. Stability analysis, grounded in Lyapunov theory, is performed to confirm the boundedness of the overall closed-loop system. Simulation and experimental results substantiate the efficacy of the proposed strategy in augmenting tracking accuracy and improving disturbance rejection. Preliminary simulation findings reveal that, compared to conventional admittance and sliding mode controllers that lack RBF integration, the proposed method achieves a reduction in root mean square tracking error by up to 95.0% (from 0.5658 m/s to 0.0285 m/s), and a decrease in maximum velocity tracking error by 51.7% (from 0.8552 m/s to 0.4132 m/s). Moreover, the system recovers to the desired state within 0.08 seconds, while the baseline method fails to stabilize within a 5-second simulation interval. These results highlight the superior disturbance rejection and rapid recovery capabilities inherent in the proposed RBF-enhanced control strategy. Collectively, these findings suggest that the proposed approach holds significant promise for ensuring reliable and precise rehabilitation motions within nonlinear and uncertain environments.

Keywords—admittance control, Radial Basis Function (RBF) neural network, sliding mode control, adaptive control, rehabilitation robot

Cite: I. V. Merkuryev, T. B. Duishenaliev, Guijun Wu, and Z. Z. Dotalieva, "Robust Control of a 2D Rehabilitation Robot Using Admittance and RBF Neural Network," International Journal of Mechanical Engineering and Robotics Research, Vol. 14, No. 5, pp. 511-524, 2025. doi: 10.18178/ijmerr.14.5.511-524

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).