【作者】 王鑫；

【导师】 杨光红；

【作者基本信息】 东北大学 ， 导航、制导与控制， 2016， 博士

【摘要】 随着计算机网络技术、通讯技术、自动化技术以及人工智能技术的不断发展与融合,多智能体系统协调控制问题越来越受到人们的重视,成为一个十分重要而且活跃的研究领域。由于多个简单智能体之间可以通过相互协调合作来共同完成复杂的控制目标,使得多智能体系统具有更强的鲁棒性、灵活性和可扩展性,并被广泛的应用于航空航天、分布式传感器网络、群体机器人编队以及工业生产过程等领域。由于多智能体系统具有较大的规模以及通讯传输网络的复杂性,在实际工程实践中,局部系统故障或通讯故障,将会导致全局控制目的的丧失,所带来的经济和社会损失难以估计。因此,多智能体系统对通讯网络和子系统安全性、可靠性的要求相对较高,这也使得多智能体系统协调容错控制问题成为一个具有重要理论和实际意义的研究课题。目前关于这方面的研究结果尚十分有限。本论文在总结前人工作的基础上,将多智能体系统通讯网络拓扑的结构特点与分布式自适应技术相结合,利用其各自的优点,建立了一套新的协调容错控制方法的理论框架。针对对称通讯网络结构下带有复合执行器故障的多智能体系统,通过利用直接自适应方法,提出了一套协调容错跟踪控制方案。针对切换对称扑结构下多智能体系统,通过构造拓扑依赖的多Lyapunov函数方法,给出了分布式容错跟踪控制器的设计方法。进而,针对一类非线性的多智能体系统,在切换无向拓扑的约束下,提出了一种新的可靠H∞一致性控制方案。然后,在上述研究基础上,将结果推广到具有切换有向拓扑的线性多智能体系统上,提出了一套分布式故障补偿设计方法,与已有结果相比,通过两步法的设计,解决了子系统控制增益设计与非对称的拓扑结构无法直接解耦的问题。最后,针对一类带有间歇性通讯和执行器故障的非线性多智能体系统,通过建立拓扑集合划分策略,给出了一种基于拓扑结构相关性质的自适应协调容错控制方案。文中的主要结果除了给出严格的理论证明外,还给出了实际例子和仿真,从直观的角度来说明所提出方法的有效性。全文分为八章,每一章的主要内容如下:第一、二章系统地分析和总结了多智能体系统协调容错控制研究领域的背景和发展现状,并给出与本文相关的一些预备知识。第三章主要研究带有执行器故障和外部干扰的线性多智能体系统的协调容错跟踪控制问题。假设智能体之间的通信网络拓扑为固定连通的无向拓扑。并且考虑跟随智能体系统发生复合执行器故障情况(包括执行器部分失效故障,偏移故障,中断和卡死故障)。与集中式容错控制问题相比,在多智能体网络中,只有一部分跟随智能单元统可以直接获得领航者的状态信息。通过利用邻接智能体的状态信息在线估计分布式控制器的参数,提出一种基于自适应技术的协调容错跟踪控制器的设计方法。进而,证明了所提方法不仅保证闭环系统中所有信号都一致有界,而且使得所有跟随者的状态能够渐进跟踪到领航者的状态。最后,通过仿真算例验证本章所提方法的有效性。第四章在第三章的基础上,针对一类带有匹配不确定性、匹配干扰以及子系统执行器故障的多智能体系统,研究在切换对称通讯网络下的协调容错跟踪控制问题。首先,基于跟随智能体的邻域信息,利用分布式自适应技术,设计容错控制器和相应的参数更新算法。然后,通过设计拓扑依赖的多Lyapunov函数方法结合拓扑驻留时间条件,证明了当拓扑驻留时间大于给定的阈值时,本章设计的分布式容错跟踪算法在补偿子系统故障的同时,保证闭环跟踪误差信号一致有界。最后,仿真算例验证了本章提出方法的有效性,与已有技术的比较说明了本章提出方法的优越性。第五章研究在切换对称网络拓扑条件下的多智能体系统的可靠H∞一致性控制问题。考虑智能体子系统带有执行器故障和外部扰动,基于每个智能体单元的邻域信息,结合自适应技术和分布式控制分配策略,给出分布式可靠控制器的设计结构。同时,通过利用拓扑相关平均驻留时间方法,以线性矩阵不等式(LMI)形式给出了子系统控制增益设计条件。进而,基于拓扑依赖的多Lyapunov函数方法,理论证明了本章设计的可靠一致性算法能保证闭环误差系统是一致有界,并具有H∞性能指标。最后,通过给出的两个数值算例进一步验证所提方法的有效性。第六章针对切换有向拓扑网络下的线性多智能体系统,提出一种新的基于分布式参考模型观测器的合作容错控制设计方法。通过设计虚拟的分布式参考模型观测器,提出基于两步法的容错控制设计方案。首先,利用跟随智能体的邻域信息去构造参考观测器,估计出领航者的运动状态;然后,利用估计信息和跟随智能体系统的状态信息,设计分散式容错跟踪控制器。通过两步设计方法,结合拓扑平均驻留时间条件,证明了跟随智能体在发生执行器故障的情况下仍然可以渐进跟踪领航者的状态轨迹。最后,利用一个网络化飞行器模型例子来验证本章方法的有效性。第七章,针对网络间歇性通讯与跟随智能体系统执行器故障同时发生的情况,研究一类带有李普希兹非线性的多智能体系统的协调容错跟踪控制问题。考虑领航智能体系统中带有未知的输入信号。首先,利用自适应算法估计控制器参数与领航者未知输入的上下界信息去构造分布式容错控制器,保证跟随智能体系统在补偿自身故障的同时可以匹配模拟出领航智能体非零输入信号。进而,根据拓扑网络性质提出拓扑集合划分规则,并结合多Lypunov函数方法以及代数图理论知识,证明了当通讯率大于给定阈值时,所设计的自适应容错跟踪控制算法可以保证多智能体系统实现一致跟踪的目标。仿真算例表明这章提出的协调容错控制方法的有效性。最后对全文所做的工作进行了总结,并指明了下一步研究的方向。更多还原

【Abstract】 With the rapid development and integration of computer network technology,infor-mation technology,automation technology and artificial intelligence technology,more and more attention is paid to multi-agent systems(MASs),which has been considered as one of the most important and active fields in the research.Also,due to complicated tasks can be accomplished by a group of agents with simple structure by cooperation,the coordination control for MASs has potential applications in broad areas including the aerospace,distributed sensor network,formation of mobile robots,industrial produc-tion process and so forth.Unfortunately,such complex distributed architecture of MASs is vulnerable to the local system faults and the unreliable communication networks in practical applications.The faulty components in some subsystems may cause the local performance deterioration and lead to agent instability,even affect the coordination goal achievement of agent networks.Therefore,the requirement of system security and relia-bility is relatively high for MASs,which also makes the cooperative fault-tolerant control problem for MASs become a challenging research topic with great theoretical and practi-cal significance.But,almost no results on such issues have been reported up to now.On the basis of the previous work,this dissertation,combining communication topol-ogy of multi-agent networks and distributed adaptive control technique and using their advantages,presents a systematically new framework of cooperative fault-tolerant con-trol design.For the linear MASs with actuator failures and undirected communication networks,the cooperative fault-tolerant controller is proposed via distributed adaptive technique.Moreover,for a class of uncertain MASs with acltuator failures,heterogeneous matching uncertainties,disturbances and time-varying undirected topology,a constructive cooperative adaptive fault compensation methodology is established on the basis of the multiple Lyapunov functions(MLFs)method and algebraic graph theory.Meanwhile,the distributed reliable H∞ consensus problem for a class of Lipschitzian nonlinear MASs with actuator faults and switching undirected topology is investigated.On the other hand,for the linear MASs under switching directed communication networks,the two-step de-sign approach for constructing the distributed fault-tolerant consensus protocol is devel-oped,which is applied to decouple the asymmetric network topology structure from the local controller gains design.Finally,for a class of nonlinear MASs with intermittent communications and actuator failures,a new constructive adaptive reliable coordination control methodology is established on the basis of the distributed adaptive technique,topology connectivity allocation,MLFs method and algebraic graph theory.All the main results are verified by simulation and part of the developed theories is applied to the sim-ulation studies of networked aircraft model.The main research contents are outlined as follows:Chapters 1-2 summarize and analyze the background and the development of coop--erative control for MASs.Preliminaries about the considered problems are also given.Chapter 3 studies the cooperative fault-tolerant tracking control problem for linear MASs subject to external disturbances and actuator faults including loss of effectiveness,bias,outage and stuck.The communication network is an undirected connected graph with a fixed topology.Different from the traditional centralized fault-tolerant tracking control problem,only part of the follower nodes can obtain the information of leader di-rectly.Based on the local state information of neighboring agents,a novel cooperative fault-tolerant tracking control scheme by using adaptive mechanism is proposed.By in-troducing the estimates of controller parameters driven by neighborhood tracking errors,it is shown that not only the actuator faults in the follower nodes can be compensated,but also the global tracking objective can be achieved.Furthermore,it is proved that all closed-loop signals are bounded and all follower nodes asymptotically tracking to the leader can be achieved in the presence of actuator faults and matched external disturbances mismatched parameter uncertainties.Finally,a numerical example is given to show the effectiveness of the proposed control scheme.In Chapter 4,based on the results in Chapter 3,the distributed fault-tolerant track-ing control problem for a class of high-order agent networks with switching undirected topology,heterogeneous matching uncertainties,disturbances and actuator faults is con-sidered.Based on the relative information of neighbors,a novel fault-tolerant consensus tracking protocol is designed for each follower node via distributed adaptive mechanism.By utilizing the MLFs method and algebraic graph theory,a sufficient condition for con-sensus tracking is established.Furthermore,it is proved that if the topology dwell time is larger than a positive threshold,the state of each follower node synchronizes to that of the leader with a bounded residual error.Finally,the simulations illustrate the advantages of our approaches compared with the existing results.Chapter 5 studies the problem of distributed reliable H∞ consensus control for MASs with actuator faults and switching undirected topology.The Lipschitzian nonlinearities,several types of actuator faults and exogenous disturbances are considered in subsystems.By utilizing the relative state information of neighbors,a new distributed adaptive re-liable consensus protocol is presented for actuator failure compensations in individual nodes.By applying the topology-based average dwell time approach and Lyapunov the-ory,the distributed controller design condition is given in terms of linear matrix inequal-ities(LMIs).It is shown that the proposed scheme can guarantee that the reliable H∞consensus problem is solvable in the presence actuator faults and external disturbance.Finally,two numerical examples are given to show the effectiveness of the proposed the-oretical results.Chapter 6 investigates the cooperative fault-tolerant tracking control linear net-worked agent systems with actuator faults and switching directed networks.Suppose the interaction network of all subsystems may switch among finite digraphs,which con-tains some non-identical directed spanning trees.Based only on the state information of neighboring agents,the two-step design methodology for constructing the distributed consensus protocol is developed.First,by using the neighborhood information,the dis-tributed reference model observer is designed to estimate the state information of leader agent.Next,on the basis of the estimation and states of subsystems,the decentralized adaptive fault-tolerant controller is proposed to compensate the actuator failures.Also,the closed-loop consensus tracking can still be guaranteed to asymptotically converge to zero in the presence of actuator faults and switching directed networks.Finally,an exam-ple is given to illustrate the effectiveness of the proposed method.Chapter 7 studies the adaptive reliable coordination control problem for a class of MASs under switching undirected topology in the presence of intermittent communi-cation constraints and actuator faults.A new fault-tolerant consensus tracking control scheme is presented by utilizing distributed adaptive technique.In contrast to the exist-ing results on cooperative tracking control problem against intermittent communications,no assumption on zero input of the leader is made in the discussions.By introducing the estimates to account for the unknown faulty efficiency factors and the bounds of the leader’s control input signal,the local controller parameters are updated in individual nodes.Also,a novel topology-connectivity-dependent average dwell time approach is proposed to deal with network switching jumps effectively.By combining MLFs method and algebraic graph theory,it is shown that the global tracking errors are uniformly ulti-mately bounded even in the cases of intermittent communication constraints and actuator failures.Simulation results are given to illustrate the effectiveness of the proposed coop-erative fault-tolerant control method.Finally,the results of the dissertation are summarized and further research topics are pointed out.更多还原