Event-triggered fixed-time adaptive neural formation control for underactuated ASVs with connectivity constraints and prescribed performance

DOI: https://doi.org/10.1007/s00521-023-08417-z
2023-03-15
Neural Computing and Applications
Abstract:In this paper, an event-triggered fixed-time adaptive neural network formation control method is proposed for underactuated multiple autonomous surface vessels with model uncertainties and unknown external disturbances under communication distance constraints. First, a time-varying barrier Lyapunov function is developed to obtain prescribed performances, such as formation error constraints, and avoid collisions in the communication range with distance limitations. Second, combining backstepping technology with neural networks, a fixed-time adaptive minimum learning parameter (MLP) is proposed to improve robustness against external disturbances and model uncertainties, and an adaptive law is designed to compensate for the approximation error of MLP. Third, a relative threshold-based event-triggered strategy is developed to greatly save communication resources without degrading control performance. Subsequently, Theorem analysis shows that all signals in the closed-loop system are bounded and practical fixed-time stable. Finally, the effectiveness of the proposed method is demonstrated by numerical simulations.
computer science, artificial intelligence
What problem does this paper attempt to address?