Research On The Tracking Problem Of Sensor Networks With Unknown Inputs

With the advancement and development of wireless electronic technology and network communication technology,wireless sensor network communication technology has become an essential emerging technology in today’s world,and the study of related problems has received more and more attention from researchers.However,in practical applications,most systems are subject to unknown perturbations,while the measurement information is easily lost during transmission due to factors such as the transmission capability of the system and communication link failures,resulting in the degradation of the estimation performance of the estimator.To address the problem of state estimation with packet loss and unknown input in wireless sensor networks,two distributed state estimation algorithms with unknown input and packet loss phenomena in different cases are proposed in this thesis.In the first algorithm,an optimal unbiased distributed state estimator based on the statistical properties of packet loss variables is designed.In the second algorithm,an optimal unbiased distributed state estimator based on the timestamp technique(which can obtain packet loss or not in real time)is designed.Finally,considering the existence of estimation differences among local estimators,the effect of local estimation differences among nodes is reduced by introducing a consistent term algorithm.The main research of this thesis includes:(1)For sensor network systems containing unknown inputs and packet loss,a distributed state estimation algorithm with unknown inputs based on statistical properties is proposed by exploiting the statistical properties of the packet loss variables.First,a distributed state estimator is designed based on the statistical properties of the packet loss variables.Then,by solving a Riccati equation and a Lyapunov equation,the estimated gain parameters under unbiased conditions can be obtained.Finally,the effectiveness of the proposed distributed state estimation algorithm is verified by a simulation case.(2)A distributed state estimation algorithm with unknown inputs based on the timestamp technique is proposed for the convergence problem of the Lyapunov equation to be considered in infinite time.First,the packet loss information at each moment is obtained using the timestamp technique,which sacrifices the computational complexity for the computational accuracy to design another distributed estimator.Then,the estimated gain under unbiased conditions is obtained by simply solving a Riccati equation.Finally,the effectiveness of the proposed distributed state estimation algorithm is verified by a simulation case.(3)A distributed state estimation algorithm with consistent terms is proposed to address the problem of estimation discrepancies among local estimators.Each node can pass its own a priori estimation information to its neighboring nodes,and a distributed state estimator is designed by introducing a consistent term.This distributed estimation algorithm can theoretically overcome the effect of differences in estimates between local estimators.