灵活性(工程)
可扩展性
分布式计算
计算机科学
分级控制系统
多智能体系统
发电机(电路理论)
分散系统
控制(管理)
控制工程
图层(电子)
工程类
人工智能
功率(物理)
统计
数据库
物理
化学
数学
有机化学
量子力学
作者
Xiangyu Wang,Yujing Xu,Yang Cao,Shihua Li
出处
期刊:Automatica
[Elsevier]
日期:2024-02-01
卷期号:160: 111402-111402
标识
DOI:10.1016/j.automatica.2023.111402
摘要
In this paper, a hierarchical design framework is proposed for distributed control of multi-agent systems. Different from the traditional distributed design philosophy that directly couples the agents' cooperation and individual regulations, the proposed framework decouples these regulations and separates the distributed control design into two layers, i.e., the layer of reference signal generator design and the layer of tracking controllers design. Specifically, the nodes of the generator play the roles of some virtual agents that achieve the distributed control goal and the agents' tracking controllers are used to track the reference signals produced by the generator. The novel design framework makes the agents reach the distributed control goal well and it has better design flexibility/practicality and stronger versatility/scalability. To demonstrate these advantages, two types of representative examples on hierarchical design are exhibited for multi-agent systems with mismatched disturbances and those with communication/input delays, respectively. All examples confirm that the proposed design framework is a natural choice in the presence of some complex factors, which provides not only simpler design but also more autonomy to the agents.
科研通智能强力驱动
Strongly Powered by AbleSci AI