Adaptive neural bounded formation tracking control of multiple autonomous surface vessels subject to actuator saturation and faults

Abstract

This paper introduces an adaptive neural bounded formation tracking control strategy for a fleet of autonomous surface vessels with actuator saturation nonlinearity and faults. To begin with, by using the error transformation technique, the proposed algorithm derives the controller only based on the relative distance and angles measured by local sensor, which can avoid the requirement for supplementary leader-related information. Secondly, by employing neural networks (NNs) and auxiliary dynamic system, the system uncertainty and actuator saturation nonlinearity are handled, and a NNs-based bounded control law is derived. In addition, in order to solve actuator faults such as loss of validity, hard over faults and bias faults, the fault-tolerant adaptive law is introduced to compensate the weight matrix of the neural network and the upper bound of the fault parameters, instead of dealing with each individual matrix element. This can significantly decrease the computational workload of the algorithm, facilitating its practical implementation in engineering applications. Finally, it is demonstrated that all signals of the closed-loop system are semi-globally ultimately uniformly bounded. The performance and superiority of this control strategy are validated through numerical examples in various scenarios.