Method And Implementation Of Trace Analysis For The Train Control System Based On Hadoop Platform

Train control system(TCS)is a core equipment to ensure the safety of train operation in the railway signal system,which makes requests in the safety and reliability for the TCS.The developing trend of recording the running status information of TCS is analyzed.And the possibility to develop an application for the safety and reliability verification of the TCS based on a big data platform is studied.The parallelization of trace analysis technology is researched in this thesis.Furthermore,the capacity to verify a big trace data set.in one off-line trace analysis process of trace analysis technology parallezed on a distributed platform is discussed.The trace analysis technology is successfully parallelized and implemented on a big data platform,and then,all the limiting factors are solved successfully to expand the capacity for a large-scale trace verification.The experimental results of an artificial mass trace data set verification is analyzed,and the conclusion is drawn that the large-scale trace data set handled in one verification process can cover the TCS running for a long time,which can make the tool to play a good complementary role for the safety and reliability verification of the TCS.The theories and methods of the trace analysis technology in the runtime verification are introduced in this thesis at first,then the general structure and applicable scenario of methods are presented.An off-line trace analysis tool is developed based on MapReduce computing framework on Hadoop platform,which is called trace analysis and verification tool.The parallelization of off-line runtime verification technology is implemented in this tool,which makes the tool able to process a trace data set on a distributed platform.Using the sort phase of MapReduce framework to realize the second sort makes data collection sorted.On the basis,the space occupied by the whole data set which must be stored in the memory is optimized by using an interval storage structure in the temporal logic operational stage.In addition,the further refinement of parallelization is successfully implemented by the tool,which makes the tool able to call all nodes in a cluster in theory and use fully the calculating ability of a distributed platform.To allow the tool to handle the infinite size of the trace data set regardless of the size of node memery on a distributed platform in theory,a new algorithm is used in this theis.The larger the scale of trace data set,the longer the covering time range is,which makes the result of trace verification more reliable and the tool more valuable in practice.Finally,some responsiveness properties of the "fault-to-safety principle" about the vehicle equipment of TCS are extracted from?The Technical Specification for the Vehicle Equipment of CTCS?,and some linear temporal logic formulas are constructed to satisfy these properties.The trace analysis and verification tool is used to verify the recorded trace from the vehicle equipment of the TCS,and the result finally shows that the trace analysis tool is suitable for the TCS trace verification and has the capacity to get a satisfactory result.