Event-Triggered Control And Filtering For Networked Systems With Inputs And Outputs Constrained

Networked Control Systems(NCSs),a comprehensive product emerging from the mature development of a variety of disciplines such as network communication,control science engineering and computer technology,are systems connected by communication networks for signal transmission and exchange,and their main functional components include controllers,actuators,sensors,etc.NCSs have the benefits of simple installation and low maintenance costs,reducing the complexity of control system design and making their system architecture more flexible.In addition,NCSs have many features such as modularity,integration,automation,and remoteness,and are widely used in various fields such as space exploration,modern transportation systems,industrial automation,smart grid,remote diagnosis and troubleshooting,and remote operation.However,the limited bandwidth in wireless communication can lead to packet loss and communication delay,which in turn leads to system instability.To add to the problem,NCSs generally use open networks,and the data is vulnerable to cyber attacks during transmission.Therefore,issues such as security and energy saving of NCSs have been of great interest to scholars.In this thesis,the stability analysis and control problems of networked control systems(NCSs)and delayed neural networks(DNNs)are investigated and discussed,while combining with the event-triggered mechanism(ETM),considering the influence of deception attacks(DAs),establishing Lyapunov-Krasovskii functions,linear matrix inequalities(LMIs)criterion is given.The main work of this thesis can be summarized as follow:Chapter 3 investigates the master-slave delayed neural network event-triggered synchronization control problem with inputs-constrained under deception attacks.At first,an ETM with inputs-constrained is established to avoid congestion in the network channel and to increase the effective data transmission rate.Second,a set of Bernoulli distributed random variable(RVBD)is introduced to characterize the probability of an attack on the network,and a mathematical model of DA is constructed.Then,a new closed-loop system with variable delays is constructed by modeling input-constrained ETM,DA,and DNNs in a unified structure.Finally,the sufficient conditions that make the closed-loop system asymptotically stable are derived through the LKF function and LMIs techniques,and the corresponding synchronous controller is further designed,a specific joint design scheme and explicit expressions for the parameters are given,and the effectiveness of the proposed method is verified by a numerical example.Chapter 4 focuses on the problem of hybrid-driven-based H_∞filtering security for networked control systems under random occurring deception attacks.First,a hybrid triggered mechanism consisting of time-triggered mechanism(TTM)and ETM,whose operation state is described by a set of RVBD,is used to select effective data while ensuring system performance.Second,assuming that the network is subjected to two different random occurring deception attacks(RODAs),two sets of RVBDs are used to characterize the transition pattern of DAs and whether the system is under attack or not,respectively.Then,a new multi-delay filtering system is established to jointly describe the H_∞filtering,RODAs and hybrid triggered mechanism.Finally,the stability and H_∞performance of the above incremental filtering system are discussed using the LKF function method and LMIs technique,combined with Jensen’s inequality and reciprocally convex combination lemma(RCCL),and then the H_∞filtering parameters are co-designed to give explicit expressions for the parameters,and the feasibility of the designed plan is verified with three numerical examples.Chapter 5 investigates the problem of adaptive event-triggered synchronization control of master-slave delayed Lur’e systems under outputs-constrained and random occurring deception attacks.First,a novel driven mechanism,adaptive event triggered mechanism(AETM),is used to decide whether to transmit the latest data to the network.Secondly,the output signal is restricted to reduce the probability of undesirable behaviors such as oscillation and divergence in the system;two sets of RVBDs are introduced and mathematical models are constructed to elaborate on the RODAs and whether the network is under attack,respectively.Then,DNNs is extended to Lur’e system,and combined with AETM,outputs-constrained and RODAs,a new synchronization error system is modeled.Finally,the stability of the above system is analyzed and controlled by LKF function method and LMIs technology,which in turn gives an adaptive event-triggered matrix co-design scheme and an explicit expression for the parameters,and the proposed method’s effectiveness is evidenced by a numerical example.Chapter 6 investigates the problem of adaptive event-triggered synchronization control of master-slave delayed neural networks with inputs and outputs constrained under random occurring deception attacks.First,the introduction of input-constrained AETM avoids congestion in the network channel and increases the effective data transmission rate.Second,we limit the output signals considering that an unrestricted range of signals will affect the system harshly.Then,in a unified framework,an error system model with multi-variable delays is established by jointly describing AETM,RODAs and DNNs.Finally,the LKF function method and the LMIs technique are used to derive sufficient criteria to ensure the asymptotic stability of the synchronous error system,to provide a joint design method for the synchronous controller and the adaptive event trigger matrix,to give the parameters of the design in explicit expressions,and to verify the effectiveness of the proposed method by a numerical example.