Event-Triggered Quantitative Prescribed Performance Neural Adaptive Control for Autonomous Underwater Vehicles
Yi Shi,Wei Xie,Guoqing Zhang,Weidong Zhang,Carlos Silvestre
DOI: https://doi.org/10.1109/tsmc.2024.3357252
2024-01-01
IEEE Transactions on Systems, Man, and Cybernetics: Systems
Abstract:This article proposes an event-triggered quantitative prescribed performance neural adaptive control method for autonomous underwater vehicles (AUVs). At kinematic level, to achieve a quantitative predetermined tracking performance without violating user-defined transient indices, a quantitative prescribed performance control (QPPC) scheme is devised, where the overshoot of the transient tracking response can be specified by a quantitative design relationship. To pursue a tradeoff between tracking accuracy and resource saving, a hybrid threshold-based event-triggered mechanism (HTETM) is designed and incorporated into the AUV controller design procedure. Additionally, a modified echo state neural network (MESNN) is employed for disturbance estimation, where intermittent system information produced by the HTETM is used for online learning, resulting in that both the communication data throughput between the controller and actuators and the online computational load can be diminished synchronously. Finally, a control law is devised at dynamic level to compensate for the triggered error induced by the aperiodic sampling of HTETM. Simulation results are provided and analyzed to validate the effectiveness of the proposed control strategy with application to an omni directional intelligent navigator.
automation & control systems,computer science, cybernetics