Event-triggered distributed coordinated control of networked autonomous surface vehicles subject to fully unknown kinetics via concurrent-learning-based neural predictor

Yibo Zhang,Dan Wang,Yong Yin,Zhouhua Peng

DOI: https://doi.org/10.1016/j.oceaneng.2021.108966

IF: 5

2021-01-01

Ocean Engineering

Abstract:This paper is concerned with distributed coordinated control of networked autonomous surface vehicles (ASVs) guided by multiple parameterized paths. The model of each ASV is subject to fully unknown kinetics, including environment disturbances, model uncertainties, and unknown input coefficients. A coordinated control method is presented for multiple ASVs to achieve a containment formation based on event-triggered mechanisms. Specifically, an event-triggered kinematic control law is firstly developed based on aperiodic communication. The desired path during the triggering interval is predicted by the last event-triggered velocity. Then, an eventtriggered kinetic control law is developed, a third-order linear tracking differentiator is introduced to take the derivative of the kinematic control law, and the control input is updated only if the predefined event is triggered. A concurrent-learning-based neural predictor is constructed by using historical data such that model uncertainties and unknown input coefficients are identified simultaneously. Next, an event-triggered path update law is constructed by using a path-maneuvering design and the aperiodic information received from the network. The proposed method ensures the input-to-state stability of the resulting closed-loop system. Finally, simulation results are provided to demonstrate the effectiveness of the proposed distributed coordinated control for networked ASVs by using event-triggered mechanisms and concurrent-learning-based neural predictors.