| With the continuous development of trains in terms of intelligence,high speed,and comfort,the design of train structures is becoming more and more complex,and the types of application of train communication networks are becoming more and more diverse.At present,relatively mature train communication buses include WTB,MVB,CAN,and so on.However,the train communication network has low bandwidth and a single networking structure,which makes it difficult to meet the transmission requirements of large amounts of data.In contrast,the Ethernet bus has the advantages of high data transmission rate,wide application range,flexible networking,high integration,low price,and easy integration with information systems.It is a future development direction to apply it to the train communication network.At the same time,modern requirements for network service quality are becoming higher and higher.Network traffic modeling and prediction is of great significance in reducing network congestion,helping to distinguish whether network performance degradation is caused by network congestion or network equipment degradation,and improving network service performance.For the research of train Ethernet communication,the main work and contributions of this thesis are as follows:(1)Establish the communication experiment platform of the special train module based on Duagon cards.According to the design scheme,the i303 module of the Duagon device is tested for communication.The tests include the single-node transceiver test,dual-node transceiver test,single-node packet loss rate test,dual-node packet loss rate test,performance degradation test in the undervoltage environment,and network congestion performance degradation test.The results show that: the data reception and transmission analysis of the Ethernet i303 module and the PC,and the test analysis of basic performance such as packet loss rate are realized.This test lays a good foundation for the next step to build a physical simulation platform containing two ETB nodes.(2)For network traffic modeling and prediction,this thesis applies Animesh Anant Sharma’s improved recurrent neural network traffic prediction model.We conduct a full experimental comparison analysis of the improved cyclic neural network with the BP neural network and support vector machine(SVM).The results show that the prediction performance of the support vector machine is better than the prediction performance of the BP neural network.The network model is better than the prediction performance of the BP neural network and SVM.(3)The EMC test of the train Ethernet Duagon in accordance with GB / T 17626.3-2016,GB / T 17626.4-2018,GB / T 17626.5-2008,GB / T 24338.4-2018 standards.The Duagon Ethernet communication equipment used in this thesis is designed reasonably and can meet the EMC test.It has strong engineering guidance significance for the evaluation of the EMC performance of the equipment.Finally,based on the previous test analysis research,the visual operation interface is designed.The key design content is network system performance testing and degradation simulation. |