| Title |
RBF Neural Network-Based Cascade Controllers for Nonlinear Friction Compensation Under Dynamic Uncertainty |
| Authors |
(Nebiyeleul Daniel Amare) ; 손영익(Young Ik Son) |
| DOI |
https://doi.org/10.5370/KIEE.2025.74.1.109 |
| Keywords |
Cascade Control; RBF Neural Network; Supervisory Control; System Uncertainty; Nonlinear Friction Compensation |
| Abstract |
The cascade structure control method remains widely used in industrial motor position applications due to its flexible hierarchical control loops. However, its ability to maintain nominal performance in the presence of uncertainties heavily relies on the chosen compensation strategy. Although the commonly implemented Disturbance Observer-Based Control (DOBC) approach is highly effective, it could fall short when addressing nonlinearities and unknown dynamics. To overcome these limitations, this paper considers two Radial Basis Function Neural Network (RBF-NN)-based approaches for nonlinear friction compensation under dynamic uncertainties. The robustness of the proposed scheme is validated via comparative simulations against conventional cascade controllers including a Reduced-Order PI Observer (ROPIO)-based scheme. Simulation results underscore the limitations of traditional linear controllers in handling dynamic disturbances and demonstrate the superior performance of RBF-NN-based controllers, with the supervisory RBF-NN outperforming the adaptive RBF-NN in terms of disturbance rejection. |