Dual-Loop Adaptive Sliding Mode and RBF Neural Network Control for Robust Trajectory Tracking of Wheeled Mobile Robots

Abstract

Trajectory tracking of wheeled mobile robots (WMRs) is significantly affected by kinematic parameter uncertainties, dynamic model uncertainties, and external disturbances, particularly under high-speed operating conditions, where conventional controllers often suffer from degraded tracking accuracy or even loss of closed-loop stability. To address these challenges, this paper proposes a dual-loop trajectory tracking control strategy that combines an adaptive sliding mode controller with a continuous reaching law in the outer loop and an RBF neural-network-based dynamic compensation controller in the inner loop. Different from existing approaches that separately address kinematic uncertainty compensation or dynamic disturbance rejection, the proposed framework establishes a hierarchical uncertainty compensation mechanism, where the outer loop handles uncertain kinematic parameters and the inner loop compensates unknown nonlinear dynamics and external disturbances. In the outer loop, the continuous reaching law is employed to suppress the chattering phenomenon while improving trajectory tracking accuracy through online estimation of unknown kinematic parameters. In the inner loop, the RBF neural network approximates unknown nonlinear dynamics and external disturbances in real time, and its adaptive weight updating law is derived from Lyapunov stability analysis to enhance velocity tracking performance and dynamic robustness. Furthermore, a unified Lyapunov-based stability analysis is developed to guarantee the boundedness of all closed-loop signals and the convergence of tracking errors to a tunable small neighborhood. Comparative simulation studies under normal, high-speed, and disturbed operating conditions are conducted against a conventional sliding mode controller (SMC) and an adaptive sliding-mode disturbance-attenuation trajectory tracking controller (ASMDATC). The results demonstrate that the proposed controller achieves smaller accumulated tracking errors, smoother wheel torque responses, and faster transient convergence under normal operating conditions. Under high-speed conditions, the proposed controller maintains stable trajectory tracking and low tracking errors, whereas ASMDATC exhibits degraded performance and the conventional SMC loses closed-loop stability due to numerical divergence. These results verify that the proposed dual-loop control strategy effectively integrates chattering suppression, uncertainty compensation, and robust trajectory tracking for high-speed WMR applications.