Research Of Data-driven-based Test Method For FAO System

In recent years, with rapid development of urban rail transit business, the FAO system arises at the historic moment by virtue of its highly automated control mode. The driving mode of FAO is that the highly centralized signal system completes all of the work that the driver needs to do. Comparing with the traditional CBTC system, it needs higher safety and reliability. At present, the test is one of the important means to ensure the accuracy of system functions, the purpose of test is to verify the function of the system whether meet the needs of its specification. As to FAO system, we need to seek an efficient and perfect test method to ensure its high reliability and high security.The traditional test method of CBTC system is mainly based on the simulation test method. The test flow of this method is to set up the simulation test platform at first, then write test cases according to the specification and testing purposes, at last operate the simulation test platform to perform the test case on the basis of simulation test platform. The key point of this method lies in the construction of the simulation test platform, namely the success of the simulation test platform directly affect the test result. However, the construction of the simulation test platform is a tedious process which requiring much time and effort. Besides, nowadays there is no a set of mature evaluation system for the simulation test platform, its test results has not written theoretical basis, so the method has certain hidden danger.In this paper, the FAO system is analyzed in detail. The data-driven-based test method is introduced into the FAO system test; the generation of test cases is the key point of this method. Compared with the traditional test method, this method simplifies the links of building the test platform, on the other hand, this method also improve the test efficiency and completeness. Its main processes include:(1) Operational scenario analysis and description:analyzing the operational scenario on the basis of the detailed analyze of the object under test;Then, establishing the sequence diagrams of operational scenario, which is good for scene management and subsequent modeling;(2) The system behavior modeling and model validation:on the basis of the sequential graph of the operational scenarios, building the timed automata model of the research object; and then using the model validation method based on W&A to verify the timed automata model, which ensures the correctness of the model;(3) Test case generation and test cases serialization:According to the system requirements and testing purposes, writing the observer test coverage criteria algorithm, and then using the CoVer tool to generate test cases of research object with some specific purpose.At last, serializing the test cases using the depth-first algorithm;(4) Build test platform:Building the data-driven-based FAO system test platform, and then using it to verify the effectiveness of test cases and the feasibility of the method.According to the above processes, This paper designs and realizes the data-driven-based test method for FAO system, and realizes this method with awaken dormant scenarios, the results show that data-driven-based test method is feasible and which can meet the requirements of the FAO system test. At the end of the paper, analyzing some insufficients and improvements of this paper, and discussing the next research direction of this work.