State Estimation For A Class Of Uncertain Discrete-time Complex Dynamical Networks

In the past years,the synchronization control,topology identification and fault diagnosis of complex networks are studied,which are based on the assumption that the states of the complex network are obtained.But the actual complex network node size is large and the network topology structure is complex,generally can only measure the node's partial state information.Therefore,in order to further study and analyze the dynamic characteristics of complex networks,topology identification and fault diagnosis,it is very important to estimate the state of the original network nodes by constructing the state estimator of complex network.In the actual network,because of the physical characteristics of the sensor itself and technical limitations and network congestion,network failure,network attacks and other factors,in the measurement of network signals need to consider the phenomenon of random saturation of the sensor and data packet loss phenomenon.At the same time due to the growing network size,channel bandwidth resources are often limited.Therefore,we deeply study the design problem of the state estimator for a class of uncertain discrete complex dynamical networks.The main contents of this paper and the research results are as follows:Firstly,the state estimation problem is addressed for a class of discrete complex networks with randomly occurring single packet losses and sensor saturations.The network packet loss phenomenon is described by using the random variable which satisfies the Bernoulli distribution,the saturation phenomenon of the sensor is characterized by the saturation function.The gain design criterion of the state estimator is given by the stability theory and the stochastic analysis method in the form of linear matrix inequality.Finally,the experimental simulation verifies the effectiveness of the proposed method.Secondly,a state estimation method based on event triggering for discrete complex networks with random single packet dropout is studied.Using random variables satisfying the Bernoulli distribution to describe the packet loss in the network,the designed event triggering estimator can reduce the transmission of redundant data of the communication channel and save the bandwidth resources of the network to improve the utilization rate.The gain design criterion of the state estimator is given by the stability theory and the stochastic analysis method in the form of linear matrix inequality.Through the simulation,verify that the state estimator designed in this paper can effectively estimate the state of the original network.Thirdly,the state estimation problem is addressed for a class of coupled outputs discrete-time complex networks with random topology jump and packet loss.Considering the complex network topology matrix satisfies the Markov jump,using Bernoulli distributed random variables to describe the phenomenon of packet loss network,the design of the asynchronous observer overcomes the problem that the original network topology is unknown or the random topology changes.The gain design criterion of the state estimator is given by the stability theory and the stochastic analysis method in the form of linear matrix inequality.Finally,the experimental simulation verifies the effectiveness of the proposed method.