Distributed state estimation of networked systems based on state decomposition
This paper investigates distributed state estimation of discrete-time networked systems and proposes a novel distributed state estimation approach based on state decomposition.Via decoupling the observable state components of the system,each sensor in the network can estimate the corresponding components independently,which achieves the fully distributed design of the estimation approach.To cope with the redundancy of information transmission in the network,a state prediction diffusion strategy is established,and the sensors only transmit the predictions of their observable state components outward,which significantly reduces the consumption of computational and communication resources of the nodes in the sensor network.To prove the boundedness of the error covariance of the estimation approach designed in this paper,a compact form of error covariance iterative equation is constructed,and sufficient conditions for the existence of upper and lower bounds of the covariance are obtained.Finally,the effectiveness of the designed estimation approach is verified in simulation by comparing the estimation accuracy and robustness with existing estimation approaches.
distributed state estimationnetworked systemssensor networksstate decompositionboundedness analysis
万方数据
