Research On Driver’s Fatigue Detection Based On Brain Signal

The rapid development of economy improves people’s living standards,which also has brought the prosperity of transportation.The amount of motor vehicle is booming dramatically,so that the incidence of traffic accidents is increasing,most of which are caused by fatigue driving.This paper is based on the background of the detection and warning of driver fatigue,which has carried on the related research work around the detection of driver fatigue.Research work will now be summarized as follows:(1)To the question of denoising in original EGG data,the paper proposed decompositon level 5 by sym5 function and the signal denoising method based on Bayesianestimation.By using the "FFT-RK"method,the signals are converted to frequency feature,It proposed that the sub-band signal energy of delta,alpha,beta,and theta can be used as the input vectors of the neural network,which will be combined with the artificial neural network in the recognition of brain wave signal.We have researched separately the BP improved algorithm based on the connection weight and the BP network algorithm improved by the new theory.The network simulation experiment proved that for the recognition of brain wave signal,the genetic BP neutral network algorithm is superior to other BP neural network algorithm.(2)In this article,K-means clustering algorithm is used to classify mental state in the light of the problem concerning the driver’s mental state identification.And then It can be realized respectively the drivers’ mental state of 2-5 class classification by displaying the characteristic quantity of concentration and meditation,and clustering results of original wave.Combined with the actual requirements and the experimental results proved that the effect of three kinds is better.(3)The experiment showed that the brain wave increases sharply in the blink of an eye.This paper proposed a fatigue detection algorithm based on electric eyes with normal blinking frequency as design idea.We can judge the man may be in a state of fatigue,who is not blinking an eye for many seconds.(4)This paper set up a complete driver fatigue detection system and developed the Android APP and Window applications to verify the above algorithm research respectively.Through experimental comparative study of alarm accuracy rate,the false alarm rate and missing report rate,we identified to determine the final results in an integrated three kinds of algorithm.And The results showed that the accuracy is improved.On the APP,alarm will be prompted to drivers in real-time voice and the information such as the location information and brain wave data information synchronously uploaded to the server,and then real-time monitoring personnel on the server side can also see the user’s physiological status and location information,etc.