Research On Driver Load And Risk Identification Model Suitable For Urban Road Traffic Scene

Entering the new era,with the enhancement of economic strength,the amount of vehicles is also rising year by year.How to reduce the occurrence of road traffic accidents has become the focus of the research field.For the cause of reduce the occurrence of traffic accidents and improve the driver’s driving experience,the research on the driver’s state and driving risk in the driving process is becoming more and more extensive and in-depth.At present,the research on driving state identification is usually carried out only in a single driving scene,which lacks the adaptability of the scene,and the surrounding road traffic environment is constantly changing in the actual driving process.In the research of driver real-time driving risk identification,researchers often only use vehicle data as the observation of driving risk identification,without considering the physiological characteristics of the driver.Therefore,considering the adaptability of identifying driver mental load in different scenarios,this paper studies the driver mental load and driving risk model adapted to different scenarios based on multi-source data.The major work is as described below:Firstly,in order to design driving simulation experiments under different road traffic scenarios,the variable factors considered in the experiment are road geometry,traffic flow and driver mental load.Based on the driving task,different mental loads of drivers are stimulated.A driving simulation platform based on Logitech driving simulator and UC / win road driving simulation software is built.Driver vehicle data and driver physiological data(including EEG and ECG signals)are collected during the driving simulation experiment.Other influencing variables are controlled during the experiment to ensure the effectiveness of the driving simulation experiment data.Secondly,the influence of driver mental load on driver driving behavior and physiological indexes in different road traffic scenes is analyzed.Considering that the characteristic importance and optimal identification algorithm of driver mental load may change in different scenes,a driver mental load identification model for adaptive urban road traffic scenes is proposed in this paper,The model includes a driving scene discrimination sub model and a driver load identification sub model.The driving scene discrimination sub model can quickly and accurately judge the road traffic scene.The driver load identification sub model selects the best feature subset and the best model algorithm in the scene based on the judgment of the driving scene discrimination sub model,then we can rapidly and precisely identify the driver’s mental load in the driving process.Finally,the driving risk in the driving process is identified.Firstly,the vehicle characteristics and physiological characteristics are extracted from the vehicle data and physiological data obtained from the driving simulation experiment,and the driving risk index in the driving process is designed.The driving data is K-means clustered based on the driving risk index to obtain the driving risk category label,Secondly,the effects of different factors(traffic flow,road geometry,driving risk)on driving horizontal risk and vertical risk are analyzed.Finally,a driving risk identification model is established based on vehicle data and physiological data to realize the prediction of driver driving risk in the driving process.