非线性约束下多智能体系统事件驱动控制设计与分析

The event-triggered control for multi-agent systems has become a hot topic and attracted great interests in recent years. Most of research results were based on linear protocols when designing the event-triggering mechanism and generally assume the communication without constraints. This project focuses on the methodology for the event-triggered control in multi-agent systems with sector nonlinear and quantized communication protocols. The topics of interests include the following three aspects: (1) The protocol design and stability analysis methods of the event-triggered control with sector nonlinear of relative/absolute states will be developed, and some criteria for the stability of multi-agent systems will be established. (2) The quantized communication with static and dynamic quantizer will be studied, and we aim to establish the completed consensus condition under dynamic quantization. Specifically, the dependencies of quantization parameters, event parameters and the system performance will be analyzed and characterized. (3) We investigate the design and synthesis for smart grids with the obtained theoretical results, and provide solutions to the distributed power quality managements. This research will contribute to enriching and developing the control methods for the multi-agent systems and provide theory support for the engineers to cope with nonlinear constraints and solve control problems with lower control cost.

多智能体的事件驱动控制是近些年群体智能控制的研究热点。现有研究主要针对线性协议设计相关的事件驱动控制策略，且对通讯交互无约束。本项目将基于扇区非线性和量化传输非线性的交互控制协议，研究多智能体系统的事件驱动控制策略，具体包括：(1)基于相对和绝对状态信息测量，研究扇区非线性协议下的事件驱动设计和稳定性分析问题。建立多智能体系统在非线性协议下的稳定性判据。(2)基于静态和动态的量化策略，研究量化传输下多智能体系统事件驱动问题，并探寻完全一致的稳定性条件。特别地，项目致力于研究系统性能与量化器参数、事件驱动参数之间的依赖关系。(3)将项目的理论研究成果运用到智能电网的分布式控制设计与综合，解决非线性交互下智能电网的电能质量管理问题。本项目的研究，不仅可以发展和丰富多智能体事件驱动控制理论和方法，还能够为实际工程中处理非线性约束和解决资源受限的控制问题提供理论支撑。

本项目主要研究了基于事件驱动控制的非线性多智能体系统的同步分析与控制器设计问题。项目按照拟定的研究计划执行，研究工作步步推进，进展顺利，按期完成了项目制定的研究内容、人才培养和学术交流等研究计划。具体执行情况如下：本项目针对多积分器系统，欧拉-拉格朗日系统以及Kuramoto振子网络等在不同应用场景下，设计事件触发机制和非线性交互协议，提出了静态、动态环境下群体协同控制方法，给出了事件触发和非线性耦合的多智能体系统的稳定性和性能分析方法；在具有反馈延迟、未知干扰等不利情况下，给出自适应和鲁棒的控制协议以保证多智能体系统达到一致性控制目标；针对网络通讯可能引发对网络的恶意攻击问题，设计了观测器以补偿信息的不完整性，并在博弈论的框架下研究了多智能体系统的弹性控制方法；研究成果还应用到电力系统中，通过引入图论和分布式事件触发控制方法，解决电力系统功率分配问题。