具有不匹配干扰的异构多智能体系统的协同控制与优化

In the complex environment, the dynamical characteristics of multi-agent systems composed of heterogeneous individuals, should be described with different kinetic equations. In the running process, since the systems are affected by multiple disturbances (such as parameters uncertainties, quantized errors, packet dropouts, communication constraints, external disturbances, etc.), mismatched disturbances may appear in the dynamic models (where the disturbances may be in the different channel with the control input). In this project, we will carry out the research on coordinated control and optimization for complex multi-agent systems with mismatched disturbances. By analyzing the dynamic characteristics of complex networks, the dynamics models of heterogeneous multi-agent systems will be constructed. By means of the performance evaluation function, the optimal control of swarming motion for heterogeneous multi-agent systems is studied. Based on the analysis of the interference characteristics of dynamic systems, we separate the interference signals from the internal, the external and the modeling. Composite layered anti-disturbance controls are established for cooperative motion of heterogeneous multi-agent systems. By applying convex optimization theories and algorithm analysis, the optimal (or suboptimal) control is researched for the complex multi-agent systems with constraint conditions. Active disturbance observer and performance function are designed with optimal control theory. The robust stability is studied for the distributed optimization control of heterogeneous multi-agent systems with mismatched disturbances, and the cooperative optimization control theories for complex multi-agent systems are established.

在复杂的工作环境中，由异构个体组成的多智能体系统的运动特性，需要用不同的动力学方程描述。系统运行过程中，会受到多种干扰（如：参数不确定性、量化误差、数据丢包、通信约束、外部干扰等）的影响，系统的动力学模型中可能出现不匹配干扰（干扰与输入不在同一通道内）。本项目拟开展具有不匹配干扰的复杂多智能体系统的协同优化控制研究。分析复杂动态网络特性，构建异构多智能体的动力学模型；基于性能评价指标，研究异构多智能体系统群集运动的最优控制。分析动力学系统的干扰特性，对系统的内部、外部和建模的干扰信号进行分离，建立具有多源干扰的异构多智能体协同运动的复合分层抗干扰控制。应用凸优化理论和算法分析，研究具有约束的复杂多智能体系统的最优（或次优）控制。应用优化控制理论，设计干扰观测器和性能指标函数，研究具有不匹配干扰的异构多智能体系统的分布式优化控制的鲁棒稳定性，建立复杂多智能体系统的协同优化控制理论体系。

多智能体系统的协同控制，正被应用在越来越多的工业和国防等领域，包括机器人编队、智能电网、国防信息网络、传感器网络等。不同应用环境下具有复杂动态的多智能体系统的协同优化控制，现在已成为分布式系统相关研究的核心问题之一。.本项目主要研究了具有不匹配干扰的异构多智能体系统的协同优化控制。应用干扰估计与现代控制理论方法，对不同干扰进行分离建模，构造复合分层精细抗干扰控制，研究了具有多源干扰的多智能体系统的协同优化控制；应用滑膜控制和有限时间收敛理论，研究了具有匹配干扰和不匹配干扰的异构多智能体系统的编队运动；应用凸优化理论和主动抗干扰控制方法，设计保证异构多智能体系统的协同运动的最优控制器，研究了具有不匹配干扰的复杂多智能体系统协同运动的优化控制，建立了复杂多智能体系统的协同优化控制理论体系。.基于前期研究，项目团队共发表科研论文52篇，其中SCI检索20篇，EI检索20篇。这些研究成果为分布式网络系统的协同优化控制的深入应用打下了坚实的理论基础。