Research On Driver's State Detection And Application On Man-Machine Shared Driving

With the development of environmental perception and artificial intelligence,the level of driving automation is growing higher,which have already made great progress.However,it needs a long time to achieve the level of full driving automation.In the partial driving automation term,the driver and the intelligent vehicle accomplish the driving task together,namely Man-Machine Shared Driving.The Man-Machine Shared Driving is an important research field of the intelligent vehicle.Detecting the condition of the driver and assisting the driver finishing the dynamic driving task is important to traffic safety.During driving,fatigue driving and distraction can easily lead to traffic accident.To avoid this,the intelligent vehicle have to equip the ability of detecting the driver condition,combined with the environmental information,informing the driver warning signal and planning a reasonable path to avoid the obstacle when there is a potential risk.The main research work includes the following three parts:Firstly,get the image of driver through CCD camera.This paper use the ASM algorithm to extract the feature points of the driver's face.According to the feature points,fix the eye region.Compute the HOG descriptor of eye region and using SVM to estimate closure of eye.Estimate the fatigue level of driver according to PERCLOS index.Secondly,choose five feature points of the driver's face,namely the corner of eye,the cheek and the nose.This paper utilities the POSIT algorithm to calculate the pose of the driver's head.In addition,the pose of head can approximately represent the eye gaze direction.Thirdly,take the driver condition as inputs and utility the fuzzy control to adjust the parameters of Artificial Potential Field algorithm.Make the intelligent vehicle equipped the ability to avoid the obstacle according the condition of the driver.