Research On Passive Localization Algorithm For Moving Target Using Moving Multiple Stations

Passive localization of moving target using moving multiple station,which is achieved by utilizing the characteristics of the stations to only receive and not send electromagnetic wave signals,plays a significant role in many applications including radar,sensor networks and wireless communication because it has advantages such as not actively exposing itself.However,there have highly nonlinear and nonconvex mathematical relationships between the information acquired by the stations and the unknown target parameters,which greatly increase the diffculty of locating the moving target.Hence,based on the two positioning systems of without external illuminators and with external illuminators,this dissertation investigates and gives several key techniques for solving nonlinear and nonconvex problems in the passive lovalization of moving target using moving multiple stations.The main work of this dissertation can be concluded into the following four aspects:1.Existing localization methods without external illuminators have good performance when the measurement errors are small.However,these methods do not consider the error terms of the observational vectors and matrices.Aiming at this problem,a moving target localization algorithm based on semi-definite relaxation(SDR)technique are proposed.The idea of the stochastic robust least squares is firstly utilized to transform the positioning problem into a least squares problem,which have some quadratic equality constraints.Then,applying the SDR to relax the constrained problem into a semi-definite programming(SDP)problem,which is then solved effectively by the optimization toolbox.Simulation experiments show that the proposed method has better robustness than the state-of-the-art methods.2.Localization method without external illuminators to minimize the number of stations is studied.When using only time difference of arrival and frequency difference of arrival information,the traditional two-stage weighted least squares(TS WLS)algorithms introduce additional variables,which require at least N+2 stations to localization in the N-dimensional space.The Taylor-series expansion algorithm can use N+1 stations to estimate the target parameters.However,this method needs an initial value.To address these problems,a localization method with the minimum number of stations is developed in this dissertation.The first stage of the proposed method is to separate the unknown vector of the TSWLS method.According to the relationships between the intermediate variables and the target parameters,quadratic polynomials about the intermediate variables are established and solved by the root finding formula.Then we give the target estimate values.In the second stage,the taylor series expansion is applied for modifying the estimated values of the first stage,which aims to improve the localization accuracy.Simulation results show that the proposed method can solve the localization problem via only four stations.And the proposed method can achieve Cramer-Rao lower bound(CRLB)under moderate measurement error conditions when using five stations.3.Passive localization method considering sensor parameter errors without external illuminators is studied.Taking the method of minimizing the station numbers for example,detailed derivations and analyses of the impacts of not considering sensor parameter errors are given.Subsequently,an improved TSWLS method is proposed.The first stage of the proposed method is to introduce additional variables and establish a pseudo-linear matrix equation,which is solved by weighted least squares(WLS)technique.By considering the relationship between additional variables and target parameters,the second stage construsts a new matrix equation and then gives the final solution via WLS technique.Compared with the traditional TSWLS algorithm,the proposed method can directly give the final estimates without the extra operations in the second stage.Numerical simulations confirm the proposed method can achieve CRLB at small measurement errors.4.Existing passive localization methods with external illuminators require the exact knowledge of prior measurement error variances,which may not be known in practice.To solve this problem,a localization method without measurement error variances is investigated.The proposed estimator first transforms the localization problem into a SDP problem with respect to the target position only,which is then solved by optimization toolbox.By utilizing this value,we then construct and approach another SDP problem only about the target velocity.The variances of the measurement errors can be ignored in both two SDP problems.Simulation results show that the proposed method can achieve CRLB at moderate measurement errors.