Research On Channel Matching And Implementation Of Planning Simulation System Based On Digital Map

Since the development of communication technology,channel characteristics and application scenarios have changed rapidly,and network simulation planning has become more and more important when the technology is implemented.The wireless network expands to higher frequency band,larger bandwidth and three-dimensional space.The small coverage and large number of base stations make operators spend a lot of manpower and resources in the planning stage.However,the channel matching in the current network planning stage is mostly based on artificial division and manual testing.The workload is huge and the licensing cost of foreign simulation systems is very high.Therefore,it is of great practical significance to study the channel matching technology based on digital maps and combine it into the design and implementation of the network planning simulation system.This paper aims at the actual needs of operators or their design and planning departments in channel matching and simulation systems for wireless networks,designs and implements a network planning simulation system with channel matching module based on digital map.The system can simulate the signal strength of selected areas through information such as the base station and digital map,and complete the functions of base station management,map management,data import and export,simulation calculation and visualization;The system is based on the model-view-controller architecture,the front end combines with the amap application programming interface,and the geocode system is used in the simulation process to simplify the calculation complexity of the search of surrounding base stations and map information;This system refers to the channel model,which proposed by the standardization organization 3GPP and is used in simulation calculations,including line-of-sight and non-lineof-sight models for four typical scenarios.For channel matching,this paper designs a channel scene matching module based on building vector data and flat map data in the digital map.In this paper,machine learning algorithms are used to assist the judgment of channel matching,which is mainly divided into two tasks: channel scene judgment and indoor and outdoor scene judgment.For areas with building vector data,the module is designed to extract geographical features and construct a model to divide urban and suburban areas,and then complete the judgment based on the outline and height of the building.In the experiment,different algorithms were used for testing.The urban and suburban partition model constructed by random forests has an accuracy of 99.95% on the divided urban and suburban test dataset.When the model is applied to other areas,the obvious boundary between urban and suburban areas can be seen,which verifies that the design scheme can effectively perform channel matching;For areas without building vector data,the module is designed to extract geographic features using flat map data,and use the same feature vectors to complete the two judgment tasks.The accuracy of the model in the test dataset of building vector data construction labels is 95.11% and 89.92%,respectively.This verifies that the design scheme can basically replace building vector data,which is more difficult to obtain,with the flat map data to complete the channel matching task.The system uses predictive model markup language to deploy the obtained model on the simulation platform and test its accuracy reduction,which verifies the feasibility of the deployment scheme.Finally,the process of the actual application scenario was simulated in the system test,showing the integrity of the system functions and system performance.