Research of High-Speed Train Fault Diagnosis System Based on Multi-Agent Platform

With the continuous development and use of high-speed trains, the complex systems and automotive equipment brings out more diverse type of locomotive fault, and also put forward higher requirements for fault diagnosis system. Multi-Agent can take advantage of parallel distributed processing technology and the modular design concept, the complexity of the system is divided into relatively independent Agent subsystem to complete complex diagnostic problem solving through cooperation and competition between agents. Meanwhile, taking into account some of the problem areas have accumulated rich fault cases taken CBR (Case Based Reasoning) method, establish a multi-level fault-level index, narrow the scope of case retrieval, and improve the retrieval efficiency, use the nearest neighbor matching strategy to enhance the accuracy of the retrieval. Finally, realized train fault diagnosis system based on Multi-Agent platform on the JADE platform, the experiment proved that the system has collective intelligence, better robustness and fault tolerance.