Study Of NLOS Suppression Localization Methods Using TOA And RSS

In recent years,with the increasing complexity of the application scenarios in localization technology and the increasing demand for higher localization accuracy,the problem of high-precision target localization in the Non-line of Sight(NLOS)environment needs to be solved urgently.In this context,the localization technology of hybrid time of arrival(TOA)and received signal strength(RSS)measurements has attracted widespread attention,because it can provide more data under the limited number of sensors,and complement the technical characteristics of TOA and RSS,there is the ability to obtain higher localization accuracy under NLOS conditions.Therefore,this thesis systematically studies the hybrid TOA-RSS localization methods under NLOS conditions,and modifies the methods by analyzing the performance characteristics of hybrid TOA-RSS technology to improve the localization accuracy.The main work of this thesis is as follows:1.Study the performance characteristics of hybrid TOA-RSS localization.The mean square error(MSE)and cramer-rao lower bound(CRLB)of TOA,RSS,and hybrid TOARSS based on NLOS are derived that indicate the performance improvement space of hybrid TOA-RSS better than TOA or RSS.The expression of "critical distance" is sorted out that expounded when the propagation distance is far greater than or far less than the critical distance,the hybrid TOA-RSS localization should trust TOA and RSS measurements respectively.2.Aiming at the hybrid TOA-RSS localization problem under NLOS condition,the generalized trust region sub-problem(GTRS)framework and convex optimization methods are discussed.The construction process of the methods is to pre-process the localization model with balance parameters or centralization to approximate or alleviate the NLOS bias.Express the optimization problem as a maximum likelihood problem and approximate it to the weighted least square(WLS)form.Since the WLS problem is nonconvex and underdetermined,it can be solved by transforming into GTRS frameworks,semi-definite relaxation or second-order cone programming.The simulation results under different network conditions show that the GTRS and convex optimization methods described in this thesis can achieve robust and high-precision localization under NLOS conditions.3.Aiming at the critical distance characteristics of hybrid TOA-RSS,two application methods are discussed.Application 1: Different from other methods that use all TOA and RSS measurements,L-GTRS method is designed to select the measurement type based on the propagation distance.The second application uses exponential weights based on critical distance,which can improve localization performance compared to linear weights based on propagation distance.Moreover,when the observation environment is far greater than the critical distance,these two applications greatly improve the localization accuracy due to the containment of high error RSS measurements.4.Aiming at the angle information in the hybrid TOA-RSS localization system,CFWLS and A-GTRS localization methods are discussed.The CFWLS method uses the azimuth information formed by the target and fixed observation station to transform the estimation problem into polar space for solution.The A-GTRS method considers the more realistic problem of the directional source localization.In order to obtain the localization of the source,the NLOS bias,the propagation angle and the localization are estimated alternately.