Neural-network-based adaptive output-feedback formation tracking control of USVs under collision avoidance and connectivity maintenance constraints

In this paper, we study the output-feedback formation tacking control problem for a group of unmanned surface vehicles (USVs) with modeling uncertainties under communication constraints. We consider a one-to-one communication topology, in which the leading vehicle is assigned a task to track a desired trajectory and each vehicle except for the last follower (tail agent) communicates only with one leader and with one follower. We assume that the information exchange among the vehicles is limited by some given communication radius. Under the limited communication range, connectivity maintenance and collision avoidance between the leader and follower are considered in the formation tracking control design. To compensate for the modeling uncertainties, we employ neural network (NN) approximators to estimate uncertain dynamics. Based on the dynamic surface control technique, backstepping procedure, NN-based observers, tan-type barrier Lyapunov functions, and control Lyapunov synthesis, a decentralized adaptive output-feedback formation tracking controller is presented to achieve the boundedness of the signals in the closed-loop system, while guaranteeing connectivity maintenance and collision avoidance between the leader and follower during whole operation. Simulation results demonstrate the effectiveness of the proposed formation controller. (C) 2020 Elsevier B.V. All rights reserved.