具有未知非线性信息的多智能体系统的故障估计与协同容错控制

This project researches the problems of fault estimation and cooperative fault-tolerant control for multi-agent systems with unknown nonlinear information. Based on the information transfers of different agents, for the case that the Lipschitz constants of systems are unknown and the case that the nonlinear dynamics are unknown, respectively, the main contents are given as follows: 1. The adaptive fault estimation observers are designed to estimate the system states and faults, where the Lipschitz constants are not used to reduce the conservativeness of selecting the determinate Lipschitz constant on the entire domain of definition. 2. For the multi-agent systems with unknown nonlinear dynamics, based on the designed adaptive fuzzy observer with adjustable dimension, a novel fault estimation method is proposed. In the observer design process, the adaptive compensators are added to compensate the influence of estimation errors of fuzzy basis functions on estimation results. 3. Utilizing the estimation information, the output feedback based distributed cooperative fault-tolerant control protocol is designed, which ensures that the multi-agent systems can reach consensus. In this project, combining fault estimation and disturbance estimation, fault-tolerant control and anti-disturbance control, the disturbance can be estimated, rejected and attenuated, which can reduce the influence of disturbance on research results. This project can enrich the existing theories of multi-agent systems, and can provide useful references for practical applications of multi-agent systems.

本项目针对具有未知非线性信息的多智能体系统的故障估计与协同容错控制问题，基于智能体之间的信息传递，分别在系统Lipschitz常数未知以及非线性动态未知的情况下，研究以下内容：1.不使用系统Lipschitz常数的信息，设计自适应故障估计观测器，估计系统的状态和故障，降低在整体定义区间上采用确定的Lipschitz常数带来的保守性。2.在非线性动态未知的情况下，提出基于新型维数可调自适应模糊观测器的故障估计方法，在观测器中增加自适应补偿器，补偿模糊基函数的估计误差对估计结果的影响。3.利用估计的信息，设计基于输出反馈的分布式协同容错控制协议，实现多智能体系统的容错一致性。本项目将故障估计与干扰估计、容错控制与抗干扰控制有机结合起来，对干扰进行估计、抵消与抑制，降低干扰对结果的不良影响。该研究可以丰富现有的多智能体系统理论，为其实际应用提供相关的有益借鉴。

近年来，多智能体系统的分析与设计已经成为控制领域的热点问题，然而在多智能体系统中，单个智能体上的故障可能会传递到与其相邻的智能体上，甚至蔓延到整个网络，从而造成严重影响。在本项目的资助下，项目组围绕非线性多智能体系统的故障诊断与容错控制问题进行研究，取得了一系列成果：提出新型非线性故障估计观测器设计方案，估计系统状态与故障；针对非线性多智能体系统Lipschitz常数未知和非线性动态未知的情况，结合模糊逼近理论，设计具有自适应补偿机制的故障估计观测器。利用估计的故障信息，结合干扰估计、抵消与抑制，设计具有故障补偿器和干扰补偿器的协同容错控制协议，实现多智能体系统的协同容错控制。在该项目的支持下，课题组发表学术论文21篇，其中18篇发表在IEEE Transactions on Industrial Informatics，IEEE Transactions on Systems Man Cybernetics-Systems等权威SCI检索期刊上。相应成果受到国内外专家学者的认可与积极评价。