Font Size: a A A

Design And Implementation Of LTE Measurement Throughout Diagnosis System Based On Bayesian Network

Posted on:2021-10-30Degree:MasterType:Thesis
Country:ChinaCandidate:K H YangFull Text:PDF
GTID:2518306047986379Subject:Master of Engineering
Abstract/Summary:PDF Full Text Request
In recent years,with the development of digital technology and the Internet,the ability of various industries to collect and obtain big data has greatly increased.For example,a large amount of drive test data will be obtained during the LTE network engineering optimization and daily optimization process,which brings great challenges to the analysis and diagnosis of drive test data.However,the analysis of this data relies on manual analysis by engineers,which is inefficient and poor in batch delivery,and due to the different analysis capabilities of engineers and the data of telecom network operations,each time congestion,disconnection,and faults have different performances,involving different network abnormalities,resulting in different analysis results.Therefore,to solve this problem,it is necessary to design a drive test throughput diagnosis system to learn the unsupervised dependency relationship between network parameters from the network data,that is,to construct the causal relationship between different variables through the Bayesian network,so that the network can effectively distinguish indirect or direct dependencies and reveal the causality of variables to locate the possible causes of various types of faults.In this paper,by analyzing the business scheme of LTE network optimization and combining with the business background of the drive test end,the functional and non-functional requirements of LTE drive test throughput diagnosis system were obtained.The functional requirements of the drive test throughput diagnosis system mainly include user login function,data management function,model management function,fault diagnosis function and system management function,which mainly model requirements by use case diagrams.Non-functional requirements mainly include features such as availability and high performance.Based on the functional and non-functional requirements of the system,the overall architecture design and database design of the drive test throughput diagnosis system are obtained,and on this basis,the detailed design and implementation process of the system functions are obtained.The architecture design of the drive test throughput diagnostic system mainly includes system architecture design,network architecture design and functional architecture design.The database design of the system mainly includes the E-R model design and the data dictionary design,and gives the correspondence between the tables in the database and the meaning of each field in the table.In the process of designing and implementing the drive test throughput diagnosis system,the class diagram and the timing diagram are used to design and implement the drive test throughput diagnosis system according to the definition of each functional module of the system and its business requirements,among which the design and implementation of model management function mainly introduces how Bayesian Network is applied to the drive test diagnostic system.Finally,the drive test throughput diagnosis system is tested and verified.The test part first gives the software and hardware environment required by the drive test throughput diagnosis system during the test,and then performs functional and non-functional tests of the system in conjunction with the system test environment.In the test,the functional test mainly verifies the implementation of the system's functional modules through test cases and front pages,and focus on a certain analysis of the network data diagnostic results in the fault diagnosis functional test.The non-functional test mainly verifies availability and performance of the drive test throughput diagnostic system.Through actual system testing,it is confirmed that the drive test throughput diagnostic system meets the actual functional and non-functional requirements.
Keywords/Search Tags:LTE, Fault Diagnosis, Unsupervised Learning, Bayesian Network
PDF Full Text Request
Related items