| Locating in mobile communication network is to get the geographic location of the mobile users in talk and idle state. Locating in GSM networks aims to provide operators with the network running status, adjust system parameters and eventually optimize the entire system.This thesis presents an effective way to implement positioning in GSM networks, including a constraint-minimization-based positioning algorithm that is based on the intensity of the field and TA, a pattern-recognition-based positioning method, and the implementation of the locating system.In view of the lack of flexibility and accuracy of the existing GSM network positioning mechanisms, this paper presents more accurate and practical positioning methods on the basis of the analysis of actual signaling data.The constraint-minimization-based positioning algorithm, taking into account signal attenuation in the process of the actual communication, proposes a consistent propagation model that reflects the actual situation of each cell. Under the assumption of approximate normal distribution of TA values, the confidence interval iterative algorithm is used to derive the TA models describing TA-distance relationship in different cells. The TA models are then employed to determine the coverage of the positioning point. With the parameters adjusted by the least-squares method, the Okumura-Hata propagation model is utilized to calculate the location of the characteristic points of the adjacent cells in coverage.The pattern-recognition-based positioning algorithm is on the assumption that the measurement data collected in a period of time on the same location are similar. The SVM correlation model is used to describe the relationships between the positions and the measurement data, and to predict the locations of mobile users.The GSM locating system is implemented using C# programming language, MapXtreme 2005 and SQL Server 2005. The system provides the functions of data management, information query, mobile station locating and model management, which greatly improves the efficiency of network optimization. The correctness and effectiveness of the system are verified by the actual data in the GSM network in Huizhou, Guangdong Province. It is also demonstrated that the positioning methods proposed in this paper achieves high positioning accuracy. |