Design And Implementation Of Data Processing System For Radio Environment Map

The Radio Environment Map(REM)can visually reflect the real state of the radio environment,providing the current radio environment information and historical information for the users in the wireless network service area.It is the important method for wireless resource regulation and optimization.The data needed for the construction of the existing radio environment map is mainly collected by the special monitoring devices or sensor nodes deployed.This data collection method has the disadvantages of high nodes deployment cost and poor scalability,so that the REM is mostly just for small-scale and specific application.In order to build a more complete REM with wide coverage,a appropriate data collection method and a radio environment data processing system with mass data processing capabilities and high performance are urgently required.Mobile crowd sensing is a new scheme for radio environment data in massive collection.In order to process the massive radio environment data collected by mobile crowd sensing,this thesis designs and implements a radio environment map data processing system,to achieve the construction of a large-scale REM.The software architecture of the radio environment data processing system is proposed by using the hierarchical design methodology that includes the data import layer,the data processing layer and the data visualization layer,which improves the scalability of the system.Among them,the data layer is applied for uploading the raw data to the data processing layer.The data processing layer is based on the Hadoop distributed processing framework,which is responsible for the cleaning,filtering,fusion and statistical analysis of massive radio environment data.The data visualization layer is mainly responsible for calling the network map API,mapping the position information,displaying the radio environment map,so that users can visually view and manage information.In order to improve the performance of the system,the parameters of Hadoop cluster are optimized in this thesis.For the problem of low efficiency of the storage of massive radio environment data small files in Hadoop,this thesis proposes a mass small-files storage optimization scheme based on location information and verifies its effectiveness.At present,the basic function of the radio environment map data processing system has been developed.The system test results show that the system can realize the mass radio environment data processing and visualization,and the stability of the system has reached the design goal.