Single Station Localization Technology In Macrocell NLOS Environment

With the rapid development of the wireless communication,people's demand for location-based service based on mobile stations is gradually increasing.Many scholars have proposed several localization technologies by using terrestrial wireless communication system.Although the direct wave based algorithm can obtain ideal location accuracy under LOS(Line-of-Sight)condition in the past,it gradually fails to work due to the complex terrestrial communication environment.Then the positioning error suppression algorithm in NLOS(Non-Line-of-Sight)propagation environment has become a research hotspot.The existing positioning methods in the NLOS environment are mainly divided into the following two types: One way is to limit the location of the target through multi-BS(Base Station)information,so as to achieve higher positioning accuracy.Another way is to identify the NLOS propagation and then eliminate its measurements.However,these methods face several defects.First,the joint positioning of multi-BS requires precise time synchronization,and there is a problem of signal testability of non-serving base stations.Second,incomplete detection of NLOS error always results in poor positioning accuracy.Based on this,a single station localization method using NLOS information directly is proposed in this thesis.The proposed method uses the NLOS path as an effective localization information to avoid complex NLOS detection,and can achieve a single station localization based on abundant parameter information of scattering models.Firstly,the matching algorithm of cellular network scattering model is carried out.The parameters of NLOS paths of three kinds of macrocell scattering models(i.e.ring of scattering,disk of scattering and Gaussian scatter density model)are analyzed.Then the reference and test templates for model matching are constructed.After that,the fast Earth Mover's Distance(EMD)algorithm is utilized to realize the model matching.The simulation results show that the proposed method can achieve a high matching accuracy.Secondly,relying upon the characteristics of ring of scattering,the B-LM based single station localization technology is proposed.This method builds a localization model based on the geometric position relation between the base station,the mobile station and scatterers,and uses the parameters of NLOS multipath to constrain the target's position.Then,to solve the convergence problems of traditional least square algorithm,the B-LM algorithm is put forward which can converge to the optimal target's location quickly.The theoretical analysis and simulation results show that the proposed single station localization algorithm can achieve a high positioning accuracy.Finally,a more universal localization method is developed according to the characteristics of the disk of scattering.Firstly,the grid points are determined within target's feasible region.Then,the hypothesis test is used to select the grid points satisfying the characteristic of the disk as target's candidate points.Finally the K Nearest Neighbor(KNN)algorithm is used to determine the final target's position.The experimental results demonstrate that the proposed method can obtain higher positioning accuracy while ensuring the lower algorithm complexity.