Cooperative Target Enclosing of Ring-Networked Underactuated Autonomous Surface Vehicles Based on Data-Driven Fuzzy Predictors and Extended State Observers

This article addresses the cooperative target enclosing problem of ring-networked underactuated autonomous surface vehicles (ASVs). The target velocity is unavailable, and the ASVs are subject to sideslip effects, unknown control gains, and uncertain kinetics. The control objective is to drive a fleet of ASVs to surround a moving target at a desired range and maintain a spaced formation. An integrated distributed guidance and model-free control method is presented based on extended state observers (ESOs) and a data-driven fuzzy predictor. Specifically, by using two ESOs to estimate the unknown relative kinematics induced by the unknown target velocity and unknown sideslip and a distributed target estimator to recover the target position, intermediate range keeping and phase keeping guidance laws are designed to achieve a circular motion and an evenly spaced formation, respectively. Next, a model-free fuzzy control law is developed based on a data-driven fuzzy predictor, which learns the unknown control gains and uncertain kinetics simultaneously. Finally, the closed-loop control system is proven to be input-to-state stable through Lyapunov analysis. The salient feature of the proposed method is that cooperative circumnavigating a maneuvering target with unknown velocity can be achieved without the global target information and knowledge of vehicle kinetics. Simulation results validate the effectiveness of the proposed distributed guidance and control method for cooperative target enclosing of ASVs.

Automated summary

This article addresses the cooperative target enclosing problem of ring-networked underactuated autonomous surface vehicles (ASVs) and presents an integrated distributed guidance and model-free control method. The proposed method uses extended state observers (ESOs) and a data-driven fuzzy predictor to estimate unknown relative kinematics and control gains, and designs guidance laws and control laws for target enclosing. Simulation results validate the effectiveness of the proposed method.