Research On Real-time Performance Optimization Of Train Communication Network Based On Switched Ethernet

With the development of train to high speed,heavy haul and intelligence,high capacity,high real-time performance and reliability have become the development direction of train communication network.Therefore,the research on real-time performance optimization of train communication network is of great practical significance and engineering value.In order to solve the problems of low bandwidth,low real-time performance and low reliability of the existing train communication network,the main research accomplishments are as follow:Firstly,the technical characteristics of Industrial Ethernet were analyzed,and a three-layer topology of train communication network based on switched Ethernet was proposed.Based on the redundant train backbone network,the link aggregation model,the data link layer and the network layer,the node of the redundant backbone network and its data transmission management function were designed to meet the requirements of high-capacity communication data transmission.Secondly,a real-time train communication network topology optimization scheme was proposed.By optimizing the configuration of onboard equipment,the data communication between subnets was reduced and the real-time performance of data transmission was improved;An improved adaptive genetic algorithm was proposed.By optimizing the crossover probability and mutation probability with sigmoid function,the evolutionary individuals could be improved and the adaptive function could converge rapidly.The example shows that the topology optimization scheme can significantly reduce the total amount of communication data between subnets and improve the real-time performance of data communication.Finally,a priority-based hierarchical scheduling algorithm for train communication network was proposed.The static priority was assigned according to train communication data type and real-time requirement,with the ethernet frame encapsulated.Then the frame data was scheduled by using the Earliest Deadline First Algorithm as dynamic algorithm.Afterwards the final priority was determined by combining the static priority and the dynamic algorithm.The feasibility of the algorithm was verified by network calculus theory,OPNET simulation and experiment.The results show that the hierarchical scheduling algorithm,compared with the static priority scheduling algorithm,can improve the real-time performance of high priority supervisory data and process data by over 60%respectively,which supplies a theoretical and practical reference for realtime performance of train communication network.