关键词: 多智能体系统, 一致性, 随机切换, 全分布式, 自适应控制, 输入饱和

Abstract:

This paper investigates the fully distributed scaled consensus problem of switched multi-agent systems， in which the Multi-Agent System （MAS） involves both the switching characteristics of individual agent models and communication networks. Control inputs are subject to saturation constraints， and system states are affected by scaling properties. A stochastic switching law that integrates the advantages of Markov switching and dwell-time switching is employed. The difficulties caused by random switching of agent models and communication networks， unknown global information of the communication network， input saturation constraints， and scaling properties among individual states make the fully distributed scaled consensus problem difficult to be solved directly by existing approaches. Therefore， this paper designs a novel adaptive scaled consensus algorithm and a Lyapunov function， and introduces an algebraic Riccati equation. On this basis， the conditions are established under which the switched multi-agent system can achieve almost surely exponential scaled consensus in a fully distributed fashion. In particular， the obtained conditions admit feasible solutions and are very easy to solve. Finally， an unmanned aerial vehicle swarm is introduced to verify the effectiveness of the proposed control scheme.

Key words: multi-agent system, consensus, stochastic switching, fully distributed, adaptive control, input saturation