Pseudo-target Dynamic Feasible Region Constraint Location Method Using Single Observer

Wireless location techniques estimate the target position through measurement results of signals from the target. Wireless location techniques are divided into two types by the number of observers: single-observer positioning techniques and multi-observer positioning techniques. Compared with the location techniques using multi-observer, single-observer positioning techniques need less equipment and cost. But realization of single-observer positioning techniques is difficult because the information measured by the single observer is much less than the multi-observer. On the other hand, Non-Line of Sight (NLOS) propagation exists in the complicated wireless environment, which attracts the research of mitigating NLOS errors in the field of wireless location research. The maneuvering observer can achieve more information to the positioning system, which also leads to the NLOS effect becoming even worse. As a result, there are less research results in the single-observer positioning techniques. This dissertation focuses on the single-observer positioning techniques in NLOS environment.First the dissertation addresses NLOS models and single-observer positioning techniques, aiming at building an abstract algorithm model of single-observer positioning, which can be applied in the NLOS environment. The proposed model is applicable to different NLOS error distribution models and different guide lines of algorithm. Then the instantiation of concrete NLOS error distribution model and optimization standard is studied in detail in the dissertation.1. A novel location method using single observer is proposed, which is named by pseudo-target dynamic feasible region constraint (PTDFRC) method. Algorithm models of several key problems are built. PTDFRC method regards the scatters in NLOS environment as pseudo-target, and constructs the dynamic feasible region of real target with maneuvering of the observer utilizing the information of pseudo-target. In the feasible region, the optimization methods are applied to estimate the location of real target. The key issues of PTDFRC method include feasible region construction of the real target, optimal route layout of the maneuvering observer, estimation algorithm of key parameters of NLOS models.2. In the instantiation research of PTDFRC, geometrically based single bounce circular model (GBSBCM) is selected as NLOS error distribution model. The key issues of PTDFRC method are studied thoroughly, including construction of feasible region of real target, optimal observation route layout, estimation of key parameters and performance analysis.? Angle restriction method and distribution restriction method are proposed to construct the feasible region of the real target. The angle restriction method utilizes the maximum angle spread in GBSBCM, while the distribution restriction method utilizes the uniform distribution of scatters. The feasible region is reduced with more scatters and information measured by the observer.Two optimal route layout methods are proposed. The two methods are based on different optimal principles, which are maximized convergence of feasible region area and minimized geometric dilution of precision (GDOP). Two methods are compared through simulation experiments. According to the simulation results, we draw the conclusion that the layout method based on minimized GDOP should be used during the prophase of observation in order to suppress the location errors of pseudo-targets, while the method based on maximized convergence should be used during the anaphase of observation in order to shorten the search time of optimization method.An estimation algorithm of the radius of GBSBCM is proposed, which is based on the standard deviation match. The parameter of NLOS error distribution model decides the construction of feasible region, which is the radius of scatter region in the instantiation. The radius is estimated through minimized the difference between the theoretical and sample standard deviation values. Based on the principle, the estimation algorithm can get precise result, which is proved by simulation experiments.The performance of PTDFRC method is analyzed in the dissertation. The Cramer-Rao bound of pseudo-target location is derived, which reflects the influence of measurements (AOA and changing rate of AOA) location to the location accuracy. The concept of probability of no acquisition is proposed to measure the accuracy of the feasible region. Monte-Carlo experiments are carried out to compare the angle constriction method and the distribution constriction method and analyze the location performance of PTDFRC method. As proved in simulation experiments, PTDFRC method can utilize the multi-path signals to suppress NLOS errors effectively when the information measured by the observer is limited. And the optimal route layout method can maximize the information plus of system.