Research And Implementation Of Comprehensive Drive Test Analysis System Based On Fault Tree Analysis

As the number of mobile users increasing, mobile communication company's network quality is becoming worser; network problems are becoming increasingly prominent. Subjecting to the functions of existing software, the network optimization engineers can not take full advantage of drive test data to find the network fault and the reasons efficiently. Thus there is an urgent need for an intelligent drive test analysis system to correlate the existing drive test data analysis, to find wireless network fault intelligent, to position failure causes efficiently and accurately, and improve efficiency of solving failure problems.Fault tree analysis(FTA) is recognized as a good method in analyzing the causes of the complex systems failure. By establishing logic between fault and cause, the network optimization engineers can understand and grasp the main points of the fault fully. Introducing FTA into drive test data analysis of the mobile communication company's wireless network problems and the logic of cause faults, it is feasible and useful.FTA software algorithms are based on cut-set theory. Large fault tree analysis of large-scale calculation and large-range cut-set will consume a lot of computing time and memory resources. Using recursive method to change the fault tree into Binary Decision Diagrams(BDD), can optimize the analysis process and save computing resources. However, in the fault tree to BDD conversion process, we need to sort the end events first of all. The resulting number of BDD nodes mainly depends on the ranking of the input end events. At the same time, the sort of end events will affect its structural importance and probability importance, and then affect the probability of top event occurrence. In this paper, we introduce the adjacent end event first-ordering method sorted base on the heuristic A* algorithm to sort the end events of the wireless network traffic calls fault tree, and the experiments show that the algorithm is more efficient. At the same time, we compares the value of the importance of the event with BDD method of analysis with the analysis of the traditional FTA analysis, the results showed that the former have greatly improved in efficiency with the guarantee under the premise of similar results.The main research contents of this paper are as follows:(1) Establishing the wireless network traffic calls fault tree model, obtaining the failure function of fault tree model, quantizing the knowledge of the fault and event, completing the process knowledge quantifying. Adopting the recursive method to convert fault tree into a BDD. Converting the wireless network traffic calls fault tree into BDD with the recursive method.(2) Using the adjacent end event first-ordering method sorted base on the heuristic A* algorithm in the process of fault tree into BDD of sorting the end event.(3) Illustrating the methods of qualitative analysis and quantitative analysis of fault tree analysis base on BDD, designing calculation of the top event probability and end events importance degree in BDD.(4) Through the experiment, comparing the experimental results of the end events importance of the wireless network traffic calls in the traditional FTA method with in the BDD method to test whether the results of data accuracy, and to verify whether the design approach feasibility in wireless network traffic analysis of the reliability of fault(5) Designing and implementing the comprehensive drive test analysis system based on fault tree analysis. Through the system application in mobile company, improve the network optimization engineers to monitor the wireless network and solve the wireless network question efficiency.