Urban rail transit is playing an increasingly important role in people’s daily life,and people’s requirements for its safety and reliability are becoming stricter.The Automatic Train Protection system(ATP)is responsible for ensuring the safety of the system in the entire urban rail transit system.During the operation of the train,ATP interacts with track-side equipment and on-board equipment to monitor the speed of the train and determine whether the train is superpowered.And decide whether to brake it or not,it may be life-threatening if not have a properly handled.For the security-critical software,if the correctness of the system can be guaranteed at the source during the requirements analysis phase,it will save a lot of time and resources.For the vehicle engineers,they are more concerned about whether the requirements cor-rectly describe the expected scenarios.Aiming at this problem,this paper proposes a scenario-based requirements analysis method for railway control software.This method dynamically checks the correctness of the requirements of the ATP system by executing the train scenario in the requirement model.The main content of this article is as follows:Evolutionary Requirements Modeling: In this paper,we use evolutionary mod-eling to construct formal requirements specification basis the characteristics of the ATP system requirements.First,we should determine the informal natural language description of the requirement? Second,we need to determine the data structure、input and output in-formation of each requirement,namely the semi-formal description language? finally,the”pre-post” condition in the semi-formal language is transformed into Formal description,turn into the final formal description that is the requirement specification.The evolution-ary requirements modeling solves the problem that the domain experts lack the relevant technology of formal methods and the researchers lack domain expertise to accurately describe the system requirements.And ensure that the requirements specifications accu-rately describe the system functions.Scenario description and logic check: In order to be able to dynamically moni-tor the accuracy of requirements during the requirements analysis phase,this article uses CASSL(CASco Scenario Language)to describe train scenarios,it consists of scenario ac-tions、scenario dictionary and variable assignment.The scenario dictionary provides convenience for testers’ writes.The scenario actions describe the running process of the train.And the variable assignment part indicates the value of related variables during the train running process,which provides conditions for the execution of the scenario.In addi-tion,in order to ensure the rationality of the physical scenario itself and avoid unreasonable errors,we checked the reasonableness of scenarios.Optimized generation of scenarios: After determining the scenario description rules,testers can manually write the train scenario and expected output according to the grammatical rules.And verify the correctness of the requirements by comparing the ex-pected output and the scenario execution result.However,the manual writing of the sce-nario is inefficient and the coverage rate is also low,so we optimized the scenario for these problems.The scenario is optimized from the three aspects of special variable scenario,efficiency and the scenario’s quality to verify the correctness of requirements more fully,and also improves the test efficiency. |