Research On Fault Diagnosis Of Vehicle CAN Bus System Based On Artificial Neural Network

With the rapid development of vehicle electronic technology,in addition to the automobile engine,the number of control units such as transmission,chassis,steering,ABS,instrument and so on increases.The data bus can realize that the data transmitted on a data line is Shared by multiple control units,so as to realize that several control units share the signal of a sensor.Because of the safety fault-tolerant processing function,there are many vehicles using CAN bus system at present The application of fault self-diagnosis system CAN improve the maintainability,repairability and safety of the bus system.Faw Volkswagen automobile sales co.,LTD.,production models currently adopt CAN bus control system,this paper takes the faw-vw magotan B7L vehicle as the object of experimental research,on the basis of diagnosis flow first acquisition parameters related to vehicle under the normal state of five groups of data,respectively,set up a dozen different system failure,a total of 13 kinds of state parameters of the same five groups of data,and then establish and SOM neural network fault diagnosis based on BP neural network model,the original failure data for training,and constantly adjust parameters,to the optimal training effect,and verification,the verification results are correct.The comparison between the bp-lm learning method and SOM network algorithm proves that the bp-lm learning method achieves a better effect.