通信时间受限的多智能体系统一致性控制及应用

This project considers the distributed consensus problem of high-order multi-agent systems with limited communication duration and its applications in synchronization control of manipulators with flexible joints. The following two communication models, i.e., the intermittent communication and sampled-data communication models, are considered in modeling the multi-agent systems. First, a class of consensus protocols based only on the intermittently relative states of neighboring agents is given to achieve consensus in high-order linear multi-agent systems. The effects of the intermittent communication rate and the directed topology's algebraic connectivity on the achievement of consensus is clearly addressed. Then, some consensus protocols based on the sampled-data communications are designed and employed. For the multi-agent systems with external disturbances, the H∞ performance indices of the closed-loop multi-agent systems with the above-mentioned two kinds of protocols are respectively analyzed. The optimized design problem how to make the closed-loop multi-agent systems yield a better H∞ performance index is also discussed. For high-order nonlinear multi-agent systems, some sufficient conditions for achieving consensus with intermittent and sampled-data communications are respectively provided and analyzed. The influences of the key parameters including the algebraic connectivity of the communication topology, the intermittent communication rate and the sampled interval, on the achievement of consensus are then investigated. At last, the consensus protocols are extended to solve the distributed synchronization problem of multiple manipulators with flexible joints. It can be seen from the above statements that a new method for solving the consensus problem of multi-agent systems with limited communication duration will be established in this study and the research in this project has both theoretical meanings and practical applications in synchronization control of manipulators with flexible joints and the related topics.

本项目研究通信时间受限的高阶多智能体系统的一致性问题及其在多个柔性机械臂系统同步控制中的应用。建模中，主要考虑以下两类通信时间受限的信息通信模式：间歇通信和采样通信。对线性高阶多智能体系统，首先构造仅依赖于邻居智能体间歇相对状态信息的一致性协议，刻画并讨论有向拓扑的代数连通性及系统的间歇通信率对一致性的影响；进而给出基于采样通信的一致性协议的设计方法；分析上述两种通信模式下多智能体系统一致性对外部扰动的H∞性能品质，并给出相应的优化设计协议。建立间歇通信和采样通信下高阶非线性多智能体系统实现一致性的条件；阐明通信拓扑的代数连通性、间歇通信率和采样时间区间对一致性的影响。解决间歇通信和采样通信下多个柔性机械臂的分布式同步控制问题。经过本项目的研究，初步建立一个处理通信时间受限的多智能体系统一致性问题的新方法，这将不仅推动一致性理论的发展，而且将在柔性机械臂同步控制等领域有着重要的应用。

众多大规模基础设施网络，如公路交通网络，WWW，Internet等，都可以建模为多智能体系统。人们对多智能体系统的研究不仅具有理论意义而且具有很强的工程应用背景。2003年至今，多智能体系统分布式协调控制问题逐渐成了系统与控制领域的研究热点之一，其中一致性控制是多智能体系统协调控制研究中的基本科学问题之一。本项目针对通信时间受限的多智能体系统一致性协调控制问题展开研究。主要研究了同步间歇信息通信和异步间歇信息通信对多智能体系统一致性的定性和定量影响。其中，异步间歇信息通信下的多智能体系统可以等价地建模为具有切换有向通信拓扑的多智能体系统。项目实施过程中，深入研究了间歇信息通信环境下线性多智能体系统的一致性跟踪问题；探讨了连续信息通信环境下，通信拓扑分别为固定有向图和切换有向通信图的非线性网络化多智能体系统的牵制同步/一致性跟踪控制问题。进一步研究了采样信息通信下的网络化多智能体系统牵制同步/一致性跟踪问题；以及切换有向通信拓扑下线性多智能体系统的H无穷一致性控制问题。最后，给出了分布式一致性跟踪技术在源网荷电力系统功率平衡控制中的工程应用。具体地，通过将储能装置的存、放电比例作为一致性协调变量，实现了利用多个储能平滑功率波动这一控制任务。通过项目组成员联合攻关，研究工作取得了丰硕成果。首次构造出了拓扑依赖的多重Lyapunov函数，解决了如何在切换有向通信拓扑下实现多智能体系统一致性这一科学难题，该结果解决不仅具有理论意义，而且对利用多智能体系统分布式一致性理论解决工程问题提供了科学支撑。