拉普拉斯矩阵
共识
多智能体系统
计算机科学
有向图
控制理论(社会学)
简单(哲学)
图形
生成树
强连通分量
匹配(统计)
线性系统
自适应控制
职位(财务)
财产(哲学)
数学优化
数学
理论计算机科学
算法
控制(管理)
人工智能
离散数学
数学分析
经济
财务
哲学
认识论
统计
作者
Jie Mei,Wei Ren,Yongduan Song
出处
期刊:IEEE Transactions on Automatic Control
[Institute of Electrical and Electronics Engineers]
日期:2021-12-01
卷期号:66 (12): 6179-6186
被引量:40
标识
DOI:10.1109/tac.2021.3062594
摘要
Due to the nonsymmetric property of the associated Laplacian matrix and the existence of uncertainties in the agent dynamics, the leaderless consensus problem of uncertain multiagent systems under general directed graphs is challenging. Motivated by the classical model reference adaptive control, in this article, we propose a simple yet efficient scheme, called the model reference adaptive consensus, by arranging each agent a reference output to track, where the output is generated by a linear reference model with the relative state measurements as input. We consider two typical agent dynamics, namely, the general linear dynamics with matching uncertainties and the second-order dynamics with extended matching uncertainties. Different linear reference models are designed for the above uncertain agent dynamics. For the first one, dynamical consensus is achieved under a fixed directed graph containing a directed spanning tree. For the latter one, position consensus is achieved under switching uniformly jointly connected graphs. In the proposed algorithms, only the relative states are interacted among the agents and there is no requirement on the communication of virtual signals.
科研通智能强力驱动
Strongly Powered by AbleSci AI