Research On Fault Diagnosis Of Fatigue Driving Monitoring Device Based On CAN Bus

With the increasing number of cars on the road,the number of traffic accidents is increasing each year,and a large part of them are due to fatigue driving.Because the driver works continuously for a long time,his physical condition drops and he feels tired,which makes his nerves lose vigilance and leads to accidents.Therefore,many scholars have studied fatigue driving monitoring devices and achieved some results.However,there are some faults in this device.Because of the lack of relevant research theories and diagnostic methods,a lot of resources need to be spent in the fault diagnosis of this device.Therefore it is very important to study the fault diagnosis of fatigue driving monitoring device.In this thesis,the fatigue driving monitoring device is taken as the research object,the fault diagnosis theory and method are studied,the basic composition and working principle of the fatigue driving monitoring device are analyzed,and the corresponding fault diagnosis model is established by combining the optimized BP neural network and fault tree based on sparrow search algorithm,and the failure probability of the fatigue driving monitoring device,the occurrence probability of the bottom events affecting the failure and the related importance are calculated.An on-line fault diagnosis system based on CAN bus is designed,which can diagnose whether the device fails in real time,help the maintenance personnel to locate the fault point quickly and accurately,and improve the diagnosis efficiency.The main contents of this thesis are as follows:(1)Analyze the fatigue driving monitoring device,understand the basic structure and working principle,and analyze the factors that cause the failure of the device.By comparing the advantages and disadvantages of different fault diagnosis methods,this thesis chooses BP neural network and fault tree to study the fault diagnosis of the device.(2)After analyzing the device,the corresponding BP neural network fault diagnosis model is established.However,because BP neural network has some shortcomings,the weights and thresholds of the diagnosis model are optimized by combining the latest sparrow search algorithm to improve the accuracy of the model.The model is trained and tested by the collected sample data.From the diagnosis output results,it can be seen that the diagnosis model can effectively diagnose the fault of the device with good accuracy.(3)Through the BP neural network fault diagnosis model,the faulty module can be quickly diagnosed.On this basis,combined with the characteristics of fault tree analysis in the field of fault diagnosis,the fault tree model of fatigue driving monitoring device is established,and the causes of the bottom events affecting the failure of each module are analyzed in more detail.Through qualitative analysis,all minimum cut sets causing top events are obtained and structural importance is calculated.In the quantitative analysis,fuzzy mathematics theory and expert experience method are introduced to analyze and calculate the fuzzy probability and relevant importance of each bottom event.Through these data,we can more accurately analyze the influence degree of various factors leading to device failure.(4)Design an online fault diagnosis system based on CAN bus,explain the scheme and structure of the system,and introduce some hardware circuits and software flow,so as to realize the function of online fault diagnosis of fatigue driving monitoring device.