Research Of Data Fusion Scheme Of CTCS-3Based On Colored Petri Net

In order to meet the needs of railway transportation, our country develops China Train Control System level-3(CTCS-3) with independent intellectual property rights by itself. At present, CTCS-3technical standards and requirements had been completed, and the projects of high-speed railway are also under construction in the overall. CTCS-3includes advanced technology and research results, and its speed and safety can reach the high level in the world. Right now, it's an important work for rail traffic researchers that further improving the system safety and excluding system potential hazards through mature technical means and new ideas in the existing system framework. Furthermore, there is an important significance for long-term development of high-speed railway technology. Data fusion is a comprehensive technology that includes a variety of theoretical knowledge and technical means, and is extended and applied in many fields. Among them, train control systems also adopts some of the data fusion technology in the sensor and transport layers, but there is still a missing at the system and decision-making layers. And CTCS-3integrates with two different train-ground communication methods to achieve the across lines operation and degraded operation in Railway Passenger Dedicated Lines. Because of high-speed railway control system has C2part and C3part work independently, there is lack of communication and restriction mechanism between the two parts in system layer. So it has positive significance to use data fusion to further improve the safety of CTCS-3.This paper proposed the fusion scheme of train control data based on data fusion by analyzing information transmission and system structure, and used Colored Petri Net (CPN) to model and simulate this scheme. Firstly, CTCS-3and CTCS-2systems were analyzed based on data fusion technology, and system layer fusion scheme was proposed for Temporary Speed Restriction, Line Information and Movement Authority. Secondly, the scheme was modeled in order to verify the feasibility and rationality by using CPN. The system top level information model, train control data generation sub models, fault injection sub model and data fusion sub model were created through layered architecture function of CPN, and the information package timestamp was given to create the time series. On this basis, system status and behavior were simulated in the process of train control data transmission. When fault injecting, trigger mechanism of random exponential function is used to simulate real faults. Finally, the validity of data fusion was proved through theoretical derivation and combining with large amounts of experimental data. The simulation results showed that data fusion method can reduce the fault data output quantity to some extent, so it improved the safety of system.