Developing The Radio Interference Localization System With Application In Electric Wireless Private Networks

The electric wireless network is a type of communication technology for constructing the smart grids for the State Grid Corporation in China.In recent years,with the rapid development of wireless networks,radio interference in network has seriously degrade the quality of services in network.With respect to localization of extra-network interference sources,due to the wide coverage area and complex deployment environments of the electric power wireless networks,the commonly-used geometric positioning methods cannot meet the accuracy requirements.Therefore,in the practical interference trouble-shooting,it is often necessary for technicians to perform repeated in-site localization surveys to ensure accurate positioning,which consumes a lot of human and material resources.Therefore,more efficient and accurate interference location techniques and tools are urgently needed in the constructing and operating of power wireless private networks.Extra-network interference signals can be captured through frequency-sweeping and drive-testing,and the measurement data collecting from frequency-sweeping and drive-testing reflects the propagation characteristics of interference signals.Reverse tracking of signal propagation trajectories reveals the different trajectory coverage characteristics of the interference source locations from that of other locations in network.Accordingly,an interference source localization method based on learning to rank is proposed,which simulates interference source propagation trajectory,analyzes trajectory features,and takes the received signal strengths of interference sources and a ranking model to determine the source locations.The research and development work described in this thesis is as follows:(1)3D ray-tracing is used to simulate propagation trajectories of interference signals in complex and real-life environment.Firstly,a scenario-division algorithm is proposed for scenario modeling,the simulated interference sources are located at a variety of typical radio propagation areas,the spatial distribution of the interference signals from the sources are simulated by ray tracing,and the frequency-sweeping and driving-test points for these sources are generated;then,the signal measurement points for interference localization are selected from the widely distributed driving-test points with strong received signal strengths,and then reverse ray tracing from these measurement points generates the set of possible propagation trajectories of interference signals.(2)The interference signal propagation trajectory is rasterized to realize a learning-to-rank based interference localization model.Firstly,interference localization is modeled as a raster ranking problem by rasterizing the space where interference sources are located.Then,by analyzing the signal propagation trajectory coverage phenomena of the raster where the interference source is located,the raster feature indicators for interference localization are proposed.Next,the main path screening algorithm is used to exclude the redundant trajectories,and the trajectories are rasterized and conversed to raster features.Finally,a ranking based interference localization model is implemented,which takes the LambdaRank ranking approach and aims to locate the raster where the interference source is located.(3)The interference localization software system for electric wireless private networks is developed,which implements the above-mentioned positioning approach.The software modules in the system are designed in line with the micro-service architecture,and the Thrift communication framework is used to integrate the system components and support interaction among the components.The system is successfully applied to the power grid in a southern city in China.A large scale simulation and the application results in practice show that the interference positioning approaches and the localization system described in this thesis improve the interference troubleshooting efficiency and positioning accuracy in dense urban environment,and provide effective technical means and tools for the daily maintenance of electric wireless private networks.