Research On Train Operation Attitude Measuring System Based On Wireless Sensor Network

As the running speed of trains continues to increase and the vehicle equipment continues to be lightweight,the dynamics of the train itself changes during high-speed operation.When the train is running in an area with harsh natural environment,it will affect its normal operation attitude.In severe cases,there may even be a major risk of derailing and overturning the train.Obtaining the real-time running attitude of the train is of great significance for ensuring the safety and smooth running of the train.This thesis studies the train attitude measurement system from the following three aspects:(1)Wireless sensor network technology is used to measure the attitude angles when the train is running,and a train running attitude measurement system based on the wireless sensor network is designed,which mainly includes the hardware design and software program development of the system.The hardware structure design of the terminal node and the coordinator node is carried out.The terminal node mainly designs its data acquisition module,processor module,wireless communication module and power supply module and other hardware.The coordinator also aspects of the hardware design includes a wireless communication module,a data processing control module,power module and the interface module.In terms of software program,it mainly designs the coordinator node,routing node and terminal node.(2)In order to ensure the timeliness and stability of the data transmission of the train attitude measurement system and optimize the working performance of the train attitude measurement system on the train,uneven clustering and double cluster-heads routing algorithm based on improved ant colony is adopted.Introduce the idea of uneven clustering to improve the energy hotspot problem in the system,set the cluster size reasonably according to the position relationship with the sink node,so as to reduce the workload of nodes near the base station,and set up dual cluster heads in the cluster.Balance the energy consumption between cluster heads to achieve the goal of network energy consumption balance.By improving the selection strategy,heuristic factor and pheromone updating method of ant colony algorithm,it is used to optimize the data transmission path between clusters.The simulation results show that the routing algorithm proposed in this paper has lower network delay and ensures the fluency and real-time performance of data transmission in the system.And at the same time it balances the network energy consumption in the system,thereby prolongs its life cycle on the train.(3)The neural network technology is used to fuse the data collected by the system,and the non-uniform clustering structure of the double cluster heads is organically combined with the neural network topology structure,and a two-level neural network data fusion model suitable for its structure is designed.A data fusion algorithm based on non-uniform clustering and improved firefly neural network(UCR-IFABP).The firefly algorithm is improved through mutation and dynamic step update mechanism,which is used to optimize the BP neural network.The optimized initial value is further trained by the BP neural network to obtain the optimal initial value.The optimized neural network overcomes the defects of poor network convergence and easy to fall into local optimal solutions due to improper selection of initial parameters.The simulation results show that the data is processed by two-level neural network fusion,which greatly reduces the amount of redundant data transmission,improves the efficiency and accuracy of data collection,and prolongs the life of the system.