Learning-Based Distributed Containment Control for HFV Swarms Under Event-Triggered Communication

Abstract

This article proposes a learning-based distributed containment control protocol for a team of hypersonic flight vehicles (HFVs) composed by some leaders and some followers in the presence of switching event-triggered communication. In contrast with most existing results concerning HFVs, the crucial characteristics of our design lie in that all leader and follower HFVs are empowered with a distributed cooperative learning capability in the sense that neural networks weights are not required to be adapted all the time, in that all follower HFVs are retained in the convex hulls spanned by leader HFVs all the time, and in that the communication among distinct HFVs is conducted based on an event-triggered strategy that utilizes a switching threshold. More precisely, by utilizing the learned knowledge represented by constant neural networks, an experience-based distributed control protocol is further proposed without the need for online adaptation. Numerical simulations studies on a group of HFVs have been conducted to verify the effectiveness of the presented results.