Radio Propagation Model Correction And Coverage Area Prediction For LTE Network

China has fully entered the 4G era,where users can pick up their smart phone to make phone calls or access the Internet anywhere and anytime.This great convenience benefits from the support of a huge network with base stations.User experience is critical for the operators,as the quality of the mobile communication service can be vital to the operators;furthermore,high communication quality is subject to favorable signal coverage.At the same time,the coverage condition can reflect the operating status of the communication system and the reasonability of the site resource distribution.Therefore,it has always been a very important work in the mobile communication industry to determine the coverage of the community.Coverage forecasting is required prior to network deployment.The principle is to carry out the CW test by selecting typical scenes,then process the data with statistical techniques to obtain the propagation loss formula,and finally calculate the maximum coverage radius of each base station through the formula.However,with economic development,cities are changing constantly,along with the change of propagation model of the base stations.Thus,there is no reliable data on the coverage of current base stations or communities.The inability of accurately obtaining the dynamic coverage of the city's mobile communication network has brought much inconvenience to network maintenance and optimization.Based on the gridded data supporting service cooperating with Guangdong mobile communication company Zhuhai branch,this thesis explores in-depth and meticulous analysis on communication model research.The outcome has provided support for the accuracy of the project,simplified the workflow of the Zhuhai Mobile Network Optimization Division,and greatly improved the efficiency of network maintenance and optimization.The major work been done and contributions are as follows:1.Adjust through the classic model in combination with on-the-spot survey data,the propagation model of Zhuhai City was customized for the geographical environment of Zhuhai City.The experimental data was adopted from the road test routinely conducted by the network optimization department on a monthly basis.The dataset formed by merging the preprocessed data of neighboring area and the data of the main service area.The experiment outcome shows that with the introduction of neighboring data,the number of correctable base stations increased and the accuracy of the model was improved.2.The propagation model is adjusted based onmachine-learning theory.In view of the poor explanatory power,low accuracy and high correcting complexity of traditional experiential model,the propagation model was corrected bydecision tree based machine-learning methodology.Experiments show that the model obtained by XGBoost algorithm not only enjoys enormous higher accuracy in propagation loss prediction,but also can capture the features that might affect propagation effects.With stronger explanatory power,it can be adopted to guide the parameter adjustment of network optimization.3.Voronoi-based coverage estimation algorithm was proposed,and the panorama was captured,by which the network optimization can be realized effectively.The algorithm works by appointing the verified base station as benchmark and dividing the urban area into several small areas by Voronoi algorithm,where the uncorrected base station adopted the propagation model of the nearest benchmark base stations.With regard to the unpredictability of certain areas by the above dividing approach,the prediction is carried out by conducting scenario comparison against map,at the same time the spatial interpolation method was used to predict the coverage of blank geographic grid;eventually the citywide coverage panorama was obtained.With great practical value,it can not only assist the operators by getting better understanding of the coverage conditions,raising the efficiency of routine network maintenance and optimization,but also provide reference to the decision-makers for rational allocation of resources,along with scientific and effective problem-solving approach when major accidents occurred.