Cooperative deterministic learning and formation control for underactuated USVs with prescribed performance
Shude He, Chao Dong, Shi‐Lu Dai, Tao Zou
Abstract
Abstract In this article, we aim to address the problem of cooperative learning from adaptive neural formation control for a group of underactuated unmanned surface vehicles (USVs) with modeling uncertainties, where the formation errors are subject to prescribed performance constraints. A coordinate transformation is introduced to overcome the difficulties caused by off‐diagonal system matrix. Under limited communication range, the connectivity preservation as well as collision avoidance among the initial connected vehicles are achieved by guaranteeing the intervehicle distances converge to small neighborhoods of the desired distance. Meanwhile, the convergence of bearing angles to small neighborhoods of desired bearing angles avoids the possible controller singularity problem arising from underactuation and achieves the predefined formation shape. The prescribed performance constraints are imposed on the formation errors to improve the transient and steady‐state performances. Using the deterministic learning theory, the modeling uncertainties are locally accurately identified/learned by the localized radial basis function (RBF) neural networks (NNs) along the union of all vehicles' state orbits in a cooperative way. The learned knowledge is stored in constant RBF networks and is reutilized to develop experience‐based formation control protocol to improve the control performance including reduction of the computational burden. Simulations are carried out to verify the effectiveness of the proposed formation controllers.