Adaptive robust motion/force control of constrained mobile manipulators using RBF neural network

This paper presents a method for trajectory tracking of constrained mobile manipulators under both holonomic and nonholonomic constraints. An adaptive robust motion/force controller with a Radial Basis Function (RBF) neural network is proposed to improve the tracking control based on the dynamic model with external disturbances and parameter uncertainties. In this proposed control strategy, the model-based robust compensator is associated with an adaptive controller, which is based on model-free RBF network; it can effectively enhance the control performance. With the proposed method, the trajectory and constraint force can be converged asymptotically to their respective values to reach the desired levels. Neural network estimation errors, as well as boundedness on parameter uncertainties, are all strictly controlled by an updated law of adaptive control. Lyapunov stability theory and the stability of a closed-loop system are ensured by the adaptive robust controller. Finally, simulation results of the proposed approach in a comparative manner with model-based and model-free controllers are performed to demonstrate the adequate performance of adaptive robust control in tracking errors and have a faster reduction rate.