Adaptive dynamic programming path tracking control of a tractor- trailer wheeled mobile robot

Abstract

Tractor-trailer wheeled mobile robots (TTWMRs) have complex nonlinear dynamics that make their precise trajectory tracking control challenging. This paper investigates an adaptive dynamic programming (ADP) approach using a critical neural network (NN) to address the tracking control for continuous-time TTWMR that is modeled in a suitable affine form. A critical NN approximates the optimal cost function and enables adaptive tuning of the control policy. Decoupled kinematic and dynamic ADP controllers are proposed for integrated path and velocity tracking. For this purpose, tracking error signals are defined for each control loop. Theoretical analysis proves closed-loop stability and convergence. Simulations demonstrate superior tracking performance compared to previous techniques, with lower errors and control efforts. This highlights the benefits of ADP for optimizing TTWMR control despite uncertainties. The adaptive optimal control enables promising capabilities for autonomous applications.