Parallel Path Following of Under-Actuated Cyber-Physical Autonomous Surface Vehicles Based on Meta-Learning Extended State Observer

This article is concerned with the problem of parallel path-following control of an under-actuated autonomous surface vehicle (ASV) through virtual-reality interaction. A parallel control architecture is proposed for the path-following of an under-actuated ASV. Specifically, a meta-learning extended state observer is developed to construct an artificial ASV system, achieving characterization of the dynamics of the actual ASV in the cyber space. Next, a meta-learning parallel path-following control law is designed based on the artificial ASV system to ensure precise path-following regardless of diverse sea environments. Finally, the artificial ASV system and the actual ASV system are concurrently propelled by the meta-learning parallel path-following control law, enabling the attainment of synchronous mapping within cyber-physical space. The input-to-state stability of the closed-loop system is established by cascade theory and all error signals are uniformly ultimately bounded. Simulation results are given to illustrate the effectiveness of the proposed parallel path-following control method for an under-actuated ASV.