Research On Driver's Stress Response Based On Auditory Warning

Automated driving and vehicle-to-everything(V2X)technology can effectively improve the operating efficiency of the transportation system and ensure driver safety through information transmission and automated control.However,at this stage there is still a certain distance from the realization of Full Driving Automation.In this transitional stage,when a dangerous situation or unexpected situation occurs,the driver still needs to make decisions such as braking.The driving warning system,which can ensure driving safety to a certain extent,can effectively help the driver to judge and identify potential dangerous situation,and accurately carry out the next operation.Compared with visual warning,tactile warning and other methods,auditory warning is a more suitable and commonly used driving warning method.However,different methods of auditory warning will have different effects on the driver.If the auditory warning cannot accurately guide the driver to operate,it will seriously weaken the effect of the warning and even have more serious consequences.Therefore,this article will study the driver's stress response based on different auditory warning methods.This research can provide a theoretical basis for improving the design of the driving assistance system in the transition from No Driving Automation to Full Driving Automation.First of all,this article starts with the driver's stress response and analyzes its stress response process and its related influencing factors.Aiming at the driver's stress response,the preliminary indicators of electrocardiographic characteristics,eye movement characteristics and vehicle handling characteristics were selected to design the "No Warning-Beep Sound Warning-Voice Warning With Information" experimental program,and to collect different auditory warning methods,the driver Index data of various characteristics.Secondly,by processing the measured data with statistical methods,the influence rules and changing characteristics of the driver's electrocardiographic characteristics,eye movement characteristics and vehicle handling characteristics under different driving proficiency and different auditory warning methods are analyzed.The results show that different driving proficiency has significant differences on driver's stress response.Only the mean heart rate of electrocardiogram index and pupil diameter data of eye movement index are significantly different between the Beep Sound Warning and the Voice Warning With Information,the data between the “No Warning—Voice Warning With Information” groups show that there are 10 significantly different indicator data,“No Warning—Beep Sound Warnin” groups show that there are 6 significant difference indicators.Finally,taking “No Warning—Voice Warning With Information” as the research object,the correlation analysis and principal component analysis of various characteristic indexes are carried out to construct a representative index set of driver's stress response.Through the entropy method,the effectiveness of the auditory warning for drivers is evaluated and graded.And based on the representation index set,the k NN algorithm of machine learning and the LSTM algorithm of deep learning are used to build a driver's auditory warning effectiveness prediction model under the auditory warning method.The results show that the prediction effect of the k NN algorithm is superior to the LSTM algorithm in the prediction of hearing warning effectiveness value and hearing warning effectiveness level,and it is more suitable for the construction of prediction model of auditory warning effectiveness.According to the research results,three ideas are put forward for its practical application,which provides certain ideas for improving the design of auditory warning and traffic safety research.