Research On LOS/NLOS Environments Identification And NLOS Error Mitigatioin

In recent years, positioning technology has been used in civil and military areas more and more widely. But the influence of NLOS(Non-Line-of-Sight) error on the positioning accuracy has not been able to get a better solution in the Ground-Based wireless localization. Therefore, this thesis focuses on LOS/NLOS identification and NLOS error mitigation problem in a special GDOP scenery. In this scenery with less base stations, the target is outside a polygon composed of multiple base stations. In order to mitigate the adverse effect of NLOS error on positioning accuracy, the target is located by the identified LOS BSs, the performance of different algorithms in cellular scene and special GDOP scene is analyzed and compared in this thesis.Firstly, the research status of NLOS error is studied. Current algorithms can be divided into two categories:mitigating the NLOS error directly, and mitigating NLOS error by NLOS identification, this thesis focuses on the second kind, use NLOS identification firstly, then mitigating the adverse effect of NLOS error on positioning accuracy by the identified LOS BSs.Secondly, localization algorithms in NLOS Environment are studied, the characteristics of special scene are analyzed from the point of view on the number and geometric distribution of base stations respectively, and the analysis model utilized in this thesis is established. This thesis mainly focus on the RWGH algorithm, IMR algorithm and Wylie algorithm. The simulation made in cellular network scene、normal GDOP scene and special GDOP scene respectively, the limitation of these algorithms in special GDOP scenarios is verified.Thirdly, the characteristics of radio wave propagation in NLOS environment are discussed. This thesis have deeply studied two kinds of NLOS identification algorithms namely the Position Residual Test(PRT) and the algorithm based on intersection area measurement. The algorithm based on intersection area measurement is improved for the special GDOP scene. The experimental results show that the improved algorithm is more suitable for the LOS/NLOS identification in the special GDOP scene.Then, based on above research, the shortcomings of the existing algorithms are analyzed, a improved algorithm based on step by step tests is proposed in this thesis. This algorithm gets a sample by making multiple measurements, and makes NLOS identification by two steps. The NLOS elements be moved out of the sample by the Variance Test in the first step, the improved intersection area measurement algorithm to identify the NLOS BSs at the second step. The experimental results proved the step by step algorithm is better than other identification algorithms. The target is located by the identified LOS BSs, and compared with other NLOS mitigation algorithms, the effectiveness of the proposed algorithm under special GDOP scene for the improvement of localization accuracy is verified by the simulation result.Finally, the research works in this thesis are summarized, the deficiency in the works is analyzed, and the direction of future research is pointed out.