Analysis And Localization Of Unknown Outside Interferences In Mobile Communication Networks

In recent years, mobile communication networks develop rapidly,2Q3G and4G networks exist side by side, and radio propagation environments become more and more complicated. A variety of issues affecting the quality of services in mobile communication networks emerge in endlessly, such as interferences from inside and outside of networks, increasing of dropping rates, degrading of communication quality, reducing of base station coverage areas, and network congestion. Therefore, network optimization is becoming increasingly important to ensure and maintain the quality of network services. Interference troubleshooting, as a part of network optimization, becomes a difficulty on account of multiple kinds of interferences, the complex localization process and high labor costs. Three-point localization and frequency scanning on the site are two commonly used technologies for interference troubleshooting. The former’s effectivity is heavily up to propagation environments in networks. While the latter is more dependent on the experiences of network optimization technician, and the troubleshooting process is time-consuming and high-cost, especially in dense urban environments.In order to tackle the laborious troubleshooting against the unknown outside network interferences, this thesis puts forward an approach of progressive automatic analyzing and localizing, which is particularly suitable for dense urban environments. This approach has two cores, i.e. the workflow for interference localization and parallel ray tracing. Directed by the interference localization workflow, the approach uses measurement data from frequency scanning to locate the specific locations of outside network interferences quickly. Some key technologies are described, such as the localization workflow, segmentation of frequency scanning and measurement data, analysis of signal variation features, determination of interference search region, parallel ray tracing based on divide-and-conquer as well as evaluation and decision of candidate locations of interferences.The above-mentioned localization workflow is divided into four steps. Firstly, the measurement data from frequency scanning is segmented and signal mutation points are extracted from the segmented data. Then, the region to search interference is inferred by analyzing signal variation of the segmented data. Next, typical test points are chosen as signal sources for parallel ray tracing to infer the candidate locations of interference. Lastly, localization results are decided by evaluating the candidate locations of interference.If the network has been founded to be jammed by some types of unknown outside network interferences and the frequency scanning and measurement data is collected, then the locations of these unknown interferences can be positioned by the workflow. The segmentation of the interference measurement data means that the interference measurement data is segmented according to the changing conditions of received signal strength, and the segment points are the signal strength saltation points, so that in each segment the received signal strengths are monotonously changing. The search region heuristics are summarized by means of analyzing the relationship between signal strength variation and buildings distribution as well as streets directions. The segmented data is fitted to obtain the signal strength variation trends, and the search region heuristics are utilized to infer the possible regions that the unknown interferences reside in so as to narrow the interference search range. Then some specific points in the segmented data are chosen as the typical test points, such as signal strength saltation points, signal strength extreme points and parts of normal signal points. On the basis of divide-and-conquer, the interference search region is divided into multiple sub-regions to which numerous rays are emitted from typical test points in parallel. For the rays emitted, parallel ray tracing is applied to tracking the wave propagation paths from typical test points to infer candidate locations of interference further. Finally, the candidate locations are regarded as the interference sources and parallel coverage analysis from these sources is conducted. The confidence levels of the candidate locations are then evaluated comprehensively by comparing the interference coverage analysis and the frequency scanning and measurement data, and the locations with higher levels are taken as the locations of the unknown interferences.The ray tracing model takes into consideration the impact of terrain and ground objects, e.g. buildings, on the wave propagation. The path losses and received signal strength are calculated during the tracing process. Considering that ray tracing in dense urban areas is computation-intensive and there are a large number of rays to be followed, in accordance with divide-and-conquer, ray tracing from interference sources are divided into multiple smaller tracing problems in smaller regions, and the scheme of parallel programming is taken to parallelly calculate several sub-problems on multi-core or multi-CPU systems. It’s proved that this scheme effectively speeds up the tracing process by reducing the computation time costs.The above-mentioned localization approach is applied to analyzing and positioning the locations of unknown interferences in the GSM network in Guangzhou, Guangdong province, and the results demonstrate that this scheme can quickly locate the unknown interferences with higher positioning accuracy and at lower labor costs. It is proved that the proposed localization approach can improve the efficiency of interference troubleshooting in dense urban areas and has good prospects in network optimization for practical networks.