Research On Distributed State Estimation And Synchronization Control For Networked Nonlinear Systems

Since the beginning of the 21st century,control tasks in typical cases such as process industries and discrete manufacturing have become more and more complex so that simple and stand-alone control systems can no longer meet the control needs.Consequently,networked nonlinear systems incorporating network communication technology and electronic information technology were booming.The core scientific goal for efficient operation of networked nonlinear systems is to construct a method for high precision and rapid estimation of the system state and to achieve efficient synchronization control of network nodes.To achieve such goals,the study of distributed state estimation based on sensor networks and synchronization of interconnected coupled systems has received a lot of attention and there are still many open issues to be investigated.The aim of this dissertation is to investigate the problem of distributed state estimation and synchronization control of the networked systems with complex characteristics such as nonlinearity,time delay,and jumping coupling under the influence of network phenomena such as bandwidth constraint and network attacks.In order to deal with the above problems,this dissertation presents some new thoughts and methods such as periodic scheduling transmission strategy,bit rate constraint model,network and control co-design algorithm,and so on,which are mainly as follows:1.For periodic nonlinear systems with time-varying time delay,the problem of distributed state estimation for time-delay periodic nonlinear systems based on scheduling sensor networks is studied by taking the actual situation of limited sensor bandwidth and node energyinto account.A periodic scheduling strategy for nodes to send information periodically is proposed by analyzing the relationship between the data exchange pattern and node energy scheduling.By constructing a delay-dependent periodic Lyapunov function,a sufficient condition that can guarantee the system stability and H?performance in the average sense is obtained,which effectively reduces the conservativeness of the system.2.For discrete-time nonlinear systems in the framework of digital communication networks,the problem of distributed state estimation for discrete-time nonlinear systems with constrained bit rates is investigated by considering the intrinsic constraints of network bandwidth.A bit rate constraint model is introduced to reflect the network bandwidth constraint.Then,an improved coding-decoding protocol that guarantees the uniqueness of the coding-decoding procedure is proposed based on this constraint model.A sufficient condition is obtained to guarantee that the distributed state estimation error system is ultimately bounded,and three optimal problems are proposed to compute the estimator gains that satisfy different estimation performance requirements.3.For discrete time-varying Markov neural networks,the finite-horizon l2-l?synchronization control problem of time-varying Markovian neural networks under mixed-type attacks is studied by considering the mixed-type attack case of randomly occurring DoS attacks in communication channels and randomly occurring deception attacks on controllers.Two random variables are used to describe the random occurrence of the mixed attack.Then,an observer-based synchronization controller is designed to overcome the difficulty of obtaining the state of the master neural network.A finite-horizon performance index is proposed,as well as the gains of the corresponding observer and synchronize controller.4.For coupled neural networks with jumping coupling,the finite-time synchronization control problem of Markovian coupled neural networks with partial mode observation is studied in consideration of the practical limitation that the controller can only partially observe the modal information due to network failures.The more practical engineering case where the synchronization controller can only obtain partial modes is described by applying a more general hidden-Markov observation model.The finite-time synchronization bounded condition is obtained by constructing a mode-dependent Lyapunov function,which effectively reduces the conservativeness of the system.The optimal gain of a non-fragile synchronization controller capable of resisting parameter perturbations is obtained by solving an optimization problem.5.For complex dynamical networks in the framework of digital communication networks,the cluster synchronization control problem of complex dynamical networks under bit rate constraint is studied with regard to the quantitative impact of network bandwidth constraint on synchronization performance.The extent to which the network bandwidth of a complex dynamic network is constrained is described by introducing a bit rate constraint model.A sufficient condition related to the bit rate is proposed to guarantee the ultimate boundedness of the cluster synchronization error dynamics.Two optimization problems are proposed to obtain the desired gain of the controller satisfing different synchronization performance metrics.Besides,a co-design problem with controller gain and bit rate allocation protocol as design parameters is considered for the first time.