多智能体系统的鲁棒故障估计观测器的设计

The Design of Robust Fault Estimation Observer for Multi-agent Systems

摘要:

针对线性变参数多智能体系统设计了有限频域鲁棒故障估计观测器。首先，根据每个智能体的绝对可测输出和相对可测输出建立了每个节点的动力学方程，结合无向通讯拓扑图及拉普拉斯矩阵得到了多智能体系统的动力学方程，通过合适的变换对多智能体系统模型进行了解耦；然后，根据解耦后的系统动力学方程设计了故障估计观测器，并通过优化技术得到了故障估计观测器增益矩阵和优良的鲁棒性能指标；最后，通过微型飞行器纵向飞行运动的例子验证了所设计的故障估计观测器的有效性,及系统参数在一定的范围内发生变动的时候，故障估计观测器依然可以准确的估计系统所发生的故障。

Abstract:

The fault estimation observer is designed for linear parameter varying multi-agent systems in finite frequency domain. First, the dynamic equation of each node is established according to the absolute measurable output and relative measurable output of each agent. The dynamic equation of multi-agent system is obtained by combining the undirected communication topology and Laplace matrix. The multi-agent system model was decoupled through appropriate transformation. Then, the fault estimation observer is designed based on the decoupled system dynamics equation, the gain matrix of the observer and the better robust performance index are obtained by the LMI optimization technique. Finally, the simulation of longitudinal flight motion of MAV are given to demonstrate the effectiveness of the designed fault estimation observer,when the system parameters change in a certain range, the fault estimation observer can still accurately estimate the faults occurring in the system.