Fixed-time adaptive path following control for autonomous ground vehicles with unknown parameters and input saturation

Abstract

This paper proposes an adaptive fixed-time path-following control strategy for autonomous ground vehicles (AGVs) subject to unknown dynamics and actuator saturation. An FLS-based adaptive approximation scheme is first employed to compensate for lumped nonlinearities caused by unknown vehicle parameters, unmodeled dynamics, and external disturbances, without requiring exact model parameters. Based on the backstepping framework, a fixed-time controller is developed to ensure that the path-following errors converge to a small neighborhood of the origin within a settling-time bound independent of the initial tracking errors. To handle actuator saturation, a smooth nonlinear approximation is introduced, and the resulting approximation error is incorporated into the control design. Furthermore, an adaptive finite-time command filter with a dead-zone operator is proposed to avoid repeated differentiation of virtual control laws, estimate unknown derivative bounds online, and reduce unnecessary adaptation caused by small filtering errors and measurement noise. Simulation results under single-lane-change and double-lane-change maneuvers show that the proposed scheme achieves better tracking accuracy and robustness than the benchmark command-filter adaptive control method.