| As people's living standard improveing, the number of private cars, which have brought us convenience, meanwhile, bring traffic safety issues,is increasing.The fatigue driving has been one. of the main causes of caused traffic accidents, and every year about 9,000 people die of fatigue driving.It is very necessary to reserch an effective method of fatigue driving detection,which has been taken an urgent attention by all countries.The research and design of this essay based on two projects - The Project of Shanxi Province Yechnology Industrial Environment Construction—" Intelligent Monitoring Device of Motor Vehicle Driving Condition " charged by my tutor and The Project of Taiyuan Technology Industrial—"The Research and Design of Monitoring Device of Fatigue Driving State of Motor Vehicle Driver " applied by me.Previous seniors mainly focused on the scattered system to research. The monitoring software of eye fatigue driving could only run on PCs, failed to transplante to lower place machine successfully, and directed eye image with Adaboost algorithm in large scope with low percentage detection; Also some seniors used ARM7TDMI to run image processing software.Because of its slow speed, the treatment effect and system function is not strong.Their reserches lacked of stability and accuracy, failed to realize information fusion.This subject was improved based on the study of the seniors, but also added a few indirect signal of fatigue drving. To design a non-contact and non-contact, direct and indirect combination coexist plan, which ensures system high precision less wrong operation good overall performance. I mainly took the following aspects of work:Firstly, I consulted a large number of relevant material of fatigue driving, studied the current domestic and international situation, determined my plan, selected development platform,related algorithms and sensor models.Secondly, the research and derive of related algorithm. The collection of pulse signal was developed by Fourier Series,and its power spectral was analyzed with Welch algorithm; Face image was detected by Adaboost algorithm, again human eye was tracked by CamShift algorithms in the small scope. The joint signal of speed, acceleration and the lateral displacement signal,which indirect reacts driver fatigue, was analyzed by Kalman filtering algorithm. Finally the information fusion was implemented secondarily with D-S evidence,to judge drvier fatigue comprehensivly and make corresponding measures.Again, the design of hardware circuit.The processor S3C2440A with MMU, which can reach 400MHz through configuration clock, run on utu-linux operating system smoothly, besides other peripheral modules, which constitute the whole hardware circuit. Also the docking port circuit UART,USB,CAN were designed, and some corresponding code was given.Finally, the above different algorithms were validated by experiments in MATLAB environment, and the effect is good. But when the hardware circuit debugged, because of the time, the laboratory condition and myself ability, some function was still owe ideal, and the interference effect was serious, particularly in cars. Later we still need do lots experiments, perfect system function. |
PDF Full Text Request