基于事件触发机制的多智能体系统分布式协调控制研究

For multi-agent systems consisting of a multitude of low-cost individual agents, to employ the event-triggering mechanism is an effective way of overcoming many emergent constraints, like those in communication and energy consumption, computational capacity and actuation capability. Taking into account all these factors, this project will focus on the following aspects. Firstly, we will investigate the modeling for multi-agent systems with asymmetric communication links and time-varying network topologies. To study the triggering event design, the triggering behavior analysis and the stability analysis, both the event-vector-space partition approach and the auxiliary function trajectory analysis approach will be applied. Secondly, in the context of rendezvous tasks, we will use the neighboring agent state feedback and the target point determination approach to study the distributed coordination controller design. Thirdly, we will formulate the communication and energy consumption performance index based on analyzing the number of triggering times. Furthermore, we will use the minimum principle and linear matrix inequalities to study the event-triggered optimal controller design under the proposed performance index. The project aims to propose a set of effective approaches to event design and event-triggered controller synthesis of multi-agent systems with asymmetric communication links and time-varying network topologies, and then put forward several optimized controller design approaches for communication and energy consumption performance. The ultimate objective is to provide theoretical support for the applications of resource-limited multi-agent systems constituted by a multitude of low-cost individuals.

在很多应用领域中，多智能体系统均需要由大规模、低成本的智能体个体组成。将事件触发机制用于该类系统可以有效突破智能体的通信消耗、能源消耗、计算能力和驱动能力等多种限制。本项目的研究内容为：(1)针对通信非对称和拓扑时变的多智能体系统建立数学模型，并利用事件向量空间划分和辅助函数轨迹分析方法研究系统的触发事件设计、触发行为分析和稳定性分析；(2)在聚集任务背景下，利用邻居状态反馈和目标点设定法研究分布式事件触发协调控制律设计；(3)通过触发数量分析建立通信和能源消耗性能指标，并利用极小值原理和线性矩阵不等式求解法研究性能指标下的事件触发优化控制律设计。通过上述研究，本项目将提出一套有效的适用于通信非对称和拓扑时变的多智能体系统的触发事件设计和分布式协调控制律设计方法，并在此基础上提出若干种通信和能源消耗性能指标下的优化控制律设计方法，为资源受限的大规模、低成本多智能体系统的应用提供理论支持。

本项目针对带有多种不同约束条件的多智能体系统，基于事件触发控制机制，研究了事件触发多智能体系统的建模、触发事件设计、触发行为分析和稳定性分析。在此基础上，进一步研究了不同事件设计下，如组合测量、时间事件混合触发、自触发等情况下，多智能体系统的分布式事件触发协调控制律设计。同时，建立了通信和能源消耗性能指标，研究了该性能指标下的事件触发优化控制律设计。本项目的宗旨是提出一套有效的适用于多目标多约束条件下多智能体系统的触发事件设计和分布式协调控制律设计方法，为资源受限的大规模、低成本多智能体系统的应用提供理论支持。基于本项目，在国际SCI期刊上发表论文7篇，包括IEEE TAC论文1篇，Automatica论文2篇，其中1篇论文入选ESI全球高被引论文数据库。