Scattering Information Geometry Constructed Location Method Using Single Observer

Wireless location techniques are series of location estimated methods sets, which estimate the target position through measurement results of wireless 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. Single-observer positioning techniques need less equipment and cost compared with the location techniques using multi-observer, more and more researchers engage in the single-oberver positioning researching areas. In the complicated wireless environment, the Non-Line-of-Sight (NLOS) propagation exists universally between the observer and the target, and the NLOS propagation is the one of the most important factors to effect on the estimation result precision of the target position location. Because of the weak ability of location information obtaining of single-observer, it cann't mitigate the NLOS propagation error effectively. The maneuvering observer can achieve more information to the positioning system, which also leads to the NLOS effect becoming even worse. The NLOS propation error is difficult to separation from the usful location information. This dissertation focuses on the single-observer positioning techniques in NLOS environment.The NLOS models and single-observer positioning techniques are introduced in the dissertation at first, which aims at building an abstract algorithm model of single-observer positioning, which is suitable for the different NLOS distribuation models. Secondly, the instantiation of concrete NLOS error distribution model and location parameters is studied in detail in the dissertation.1. A novel location method using single observer is proposed, which is named by scattering information geometry constructed location (SIGCL) method. The algorithm uses the NLOS propagation property sufficiently, which is looked as disadvantage factor in position location to estimate the target position. The single observer position problem is transform to the problem of multi-observer position from the point of view of the scatters. Several key problems in this algorithm model are also addressed.2. In the instantiation research of SIGCL, geometrically based single bounce circular model (GBSBCM) is selected as NLOS error distribution model.The circle track fitting method is proposed to estimate the scattering information. The pseudo-target tracing basic model is build through the analysis the character of the pseudo-target movement. The track of the pseudo-target is a circle, which the centre of the circle is scatter and the radius of the circle is the scattering distance. The circle fitting least square method is used to estimate the scattering information.The extended Kalman filter of composite motion is introduce to tracing the pseudo-target track. The scattering information estimation precision is decided by the precision of pseudo target track directly. The estimation of the scattering information has great deviation compared with the true scattering information value when the measured parameter has large noise, and the extended Kalman filter method is proposed to increase the precision of the pseudo target track. The constrainted total least square (CTLS) method and covariance matrices modification (CMM) method are proposed to decrease the effection of the scatter point disturbing. The principle of those methods is the minimize of the effect of the system total noise. The CTLS eliminate the relativity between the coefficient matrix and the measurment vector. The character, which is statistical indenpendency between the estimated parameter and noise, is used in the CMM method, and the system total noise is mitigated by the modification of covariance matric.The method of manuervable target tracking track is analyzed in the dissertation. In the instantiation research of manuervable target condition, we also selected GBSBCM model as NLOS error distribution model. The model of manuervable target is the reverse model of circle track fitting, and the different of the two models is the difference of AOA parameter. The manuervable target model could equal to the circle track fitting model after the reconstruction of the difference of AOA. The Kalman filter method is used to tracing the track of the target.The coordinate shifting recursion location method and scaled gene weighted location method are introduced in the dissertation. The coordinate shifting recursion location method is suit for the LOS propagation environment, ans is the optimize estimation after several coordinate shifting. The precision of the coordinate shifting recursion location method is closed to the Cramer-Rao low boundary. The scaled gene weighted location method fits to the NLOS propagation environment, and could mitigate the effect of the NLOS error remarkably.