Mimo Mobile Communication System In A Single Base Station Location Study

Wireless Location has played an increasingly important role in modern society, with its applications in wide areas such as the traffic, security and military. Unlike multiple Base Stations(BSs) location limited by NLOS and multi-path propagation, single BS location based on machine learning methods dismisses the traditional triangular geometry calculation and is no longer limited to the signal's LOS condition. On the contrary, the complex paths are favorable to location results. This article focuses on single BS location based on support vector machine method in the MIMO mobile communication system, and resolves problems of the channel model building, combined with the communication standards analysis and channel estimation in the process of location simulation.While previous channel models for single base station location were based on distributed scatters environment, this article uses the ray tracing method to simulate the spreading channel. In the knowledge that the SCM model belongs to geometry statistical channel model, and the geometric statistical channel modeling is the reduction of ray tracing model, this article combines the ray tracing method with SCM to create a channel model for location simulation. The third chapter establishes the channel model for location algorithm simulation, which is used for ray tracing, and the examination of the delay and angle information between source and field point. The path delay and angle information of the channel make the foundation of location. The known ray-tracing channel information is used to establish random channel model, which is the object of channel estimation.The accurate channel estimation is very important to mobile station location. The traditional algorithm of signal angle estimation required the second or higher order statistics of received signal matrix, which required large amount of received data and a long-time computation. The fourth chapter presents an MIMO channel angle estimation algorithm based on phase differences, utilizing phase differences of channel impulse response between the transmitting antennas(or receiving antennas) to estimate angle of departure(or angle of arrival). Compared to the traditional subspace decomposition algorithm, the number of antennas of this algorithm is not restricted by the number of signal sources, and the algorithm has less computation.Finally, this article provides a single BS location method in MIMO mobile communication system. It firstly introduces theory of machine learning location, and then proposes the location application of support vector machine in MIMO communication system. It focuses on the theory of support vector machine location and discusses its simulation performance. By comparing whether the measured angle information takes part in location simulation, it reveals the angle of channel's great importance in location application. By comparing the localization performances of three machine learning methods, it reflects the superior performance of SVM.