Intelligent Formal Research On Level Transition Operation Scenario Of CTCS-3

Modeling and analyzing on train operation system has been one of the main problems inthe research on Railway Intelligent Transportation System. With the continuous improvementof train speed and the increasing number of trains, it is extremely urgent to guarantee thesafety of train operation. As train operation system is a safety-critical and giant-complexsystem, its every link requires accurate security analysis and verification. Among which, theanalysis on train operation scenario is one of the important means to insure its security.The modeling and simulation method based on MAS could describe the huge scalesystem though the communication between the Agents, so it is very applicable to solve suchcomplex problems in safety-critical system. But Agents could not support the developmentand implementation by software, lacking the validation and analysis on model. Colored PetriNet(CPN) is a kind of formal tool which could describe and analysis the concurrent systems.It not only has a solid mathematical foundation but also has a graphical representation, bywhich the CPN could make a accurate represent on Agents’ behavior structure. Thus we couldget a more accruate and effective description about the static structure and dynamic behaviorof complex system by the combination of CPN and Agent, and have a comprehensiveimplementation on the whole system.This paper takes the hybrid modeling method that combines Agent theory with CPN forthe research on the train level transition operation scenario in CTCS. The train level transitionoperation scenario is abstracted as MAS. The trains and RBC are abstracted as Agents. Thewireless transmission network GSM-R is abstracted as the interactive envioronment amongAgents. We put forward a kind of hierarchy structure on train-to-ground Agents. A kind ofscalable Agent which has the ability to deal with complex events and could be suitable for ourscenario model is defined, and the explanation of component unit in the Agents is given. Itshows the flow chart of the Agents’ internal reasoning mechanism. Then, we carry on thedescription for the internal reasoning process. The detailed hierarchical structure model isestablished by CPN Tools which has the function to expand the subpages, which makes thetrain-to-ground communication process, the overall behavior of MAS and the internalreasoning process visualization. Finally, we establish the scenario model state space by CPNTools, and ensure the correctness of the model through the analysis of the state space report.Through multiple simulations of the model, we analyze how retransmission intervals ofaperiodic messages affect the performance of the real-time controled train in the differentcommunication failed cases, and verify the feasibility of the model.