Adaptive Task-Space Control of Five-Bar Parallel Robot Dynamic Model with Fully Unknown Using Radial Basis Function Neural Networks for High-Precision Applications

Tran, Thanh Hai
Ngo, Thanh Quyen
Uyen, Hoang Thi Tu
Nguyen, Van Tho
Duong, Tien Đoan

Journal of Robotics and Control (JRC)
Vol. 6 No. 4 (2025)

Designing a stable and accurate controller for nonlinear systems remains a significant challenge, mainly when the system contains uncertain factors or is affected by external disturbances. This study proposes an adaptive control method based on a Radial Basis Function Neural Network (RBFNN) to effectively estimate the uncertain components in nonlinear systems. The gradient descent algorithm updates the RBFNN parameters, and the control system's stability is rigorously proven based on the Lyapunov theory. The designed controller ensures accuracy under changing conditions and can adapt to nonlinear disturbances and system fluctuations flexibly. Through 45 consecutive test cycles, the system significantly improves precision and outperforms other control methods in comparative tests. This study opens up the potential for broad application in highly uncertain nonlinear MIMO systems, thanks to the effective combination of adaptive learning ability, stability, and simple implementation structure of the proposed controller.