Research On Cellular Location Techniques Based On Information Fusion

With the deepening development of the cellular wireless communication system, especially with the large-scale popularization of 3G services in recent years, the MS (mobile station) position-based services have become one of the main value-added services for operators, so intensive researches into the cellular location technology have great significance both in theory and in practice. This paper mainly studies the information fusion-based cellular location technology, in which the focus is chiefly on such aspects as the overall program and whole structure of the cellular information fusion location technology, the location algorithm based on the integration of different location information, the mobile station location decision fusion algorithm, the location error mitigation algorithm and the evaluation optimization algorithm of the fusion location system. The main points are as follows:1. The thesis, taking the practical application of the cellular location technology as a starting point, systematically analyzes the research status of the cellular location technology and the main technical problems, studies in depth the concept, principle and common methods of the information fusion technology, investigates in detail the feasibility of applying information fusion technology to cellular location in theory, and describes the difficulties in information fusion location technology and solutions for them.2. The thesis proposes the overall design plan for the information fusion technology-based cellular location system, raises an information fusion method with the data layer fusion before decision layer fusion, constructs the whole framework for the cellular information fusion location system and puts forward an information fusion location system with parallel distributed feedback structure. Then it gives a thorough analysis of this system from its vertical structure, its location data processing procedure and its overall functional structure.3. For the data level fusion location algorithm based on multi-sensor measurement parameters, the thesis re-analyzes the issue of cellular location with the multi-objective optimization theory, improves the current location methods on the basis of the feasible location region of the mobile station and comes up with a new MS location estimation method based on the discrete fitness distributed genetic algorithm, which can not only guarantee the location accuracy but also improve the robustness of the algorithm. Then, the thesis, considering the characteristics of the multi-sensor measurement fusion location method, proposes an improved genetic algorithm based on the distributed evolution of multiple populations. This algorithm can effectively improve the diversity of individuals in target populations and in turn reduce the probability of prematuration, thus ensuring that the genetic algorithm can be applied to the solving of multi-objective optimization problems.4. For the GIS geographic information-based data layer fusion location algorithm, the thesis advances the extraction method for GIS vector data and raster data and on this foundation proposes the overall strategy for the geographic information-based data layer fusion location. For the information fusion location in macro cell environment, the thesis puts forward a MS feasible region construction method based on the GBSBCM channel model. For the information fusion location in micro cell environment, the thesis presents a new location method by using the reverse tracking of rays on the ground of priori environment information.5. As to the decision layer fusion location algorithm, the thesis at first introduces the common decision layer fusion methods in information fusion technology and then focuses on the Bayesian theory-based decision layer fusion method. On this basis, the thesis comes up with decision layer fusion structure of the cellular information fusion location method and then mainly analyzes the optimal and sub-optimal decision fusion algorithms. Secondly, the thesis raises an improved particle filter-based closed-loop feedback decision layer fusion location algorithm and then elaborates on the structure and principles of the algorithm. Finally, in view of the particle degradation and depletion of particle filters, the thesis brings forward an IUKF algorithm-based particle resampling method and simulates the performance of this method.6. NLOS errors are the key problem to be solved for the cellular network-based MS location system to improve location accuracy. The thesis at first summarizes systematically the existing NLOS error (mitigation) algorithms and then poses two NLOS error mitigation algorithms of different application models. One is based on the signal slop return-to-zero principle. It reduces the computation complexity and applies flexibly to the situation where the priori environment is unknown. The other robust NLOS error mitigation algorithm is based on Kalman filter and neural network. It trains the environment coefficients through neural network and performs better than other similar algorithms.7. The thesis studies the performance evaluation and optimization of cellular information fusion location system. At first, it describes the significance, principle and method of performance evaluation and on this basis, puts forward the information feedback structure-based performance optimization method. Then, it investigates the optimization methods for the expert system-aided information fusion system and then introduces the ideas and principles of this method. Finally, on the foundation of WCDMA system, the thesis builds a cellular information fusion location test platform, gives a detailed description of the design principles and functional structure of the platform and at last conducts a simple field test with the aid of it.