Study On Driver’s Risk Perception Under Pedestrian Crossing Conflict On Mountain Highways

The situation of mountainous highway is complex,the grade of highway is low,and traffic accidents occur frequently.Due to the large number of openings and irregularity of mountain roads,pedestrian crossing behavior is more random and difficult to predict.Drivers face greater risks in the face of the instability and uncertainty of pedestrian crossing,which leads to the high accident rate of pedestrian crossing on mountain roads.Nowadays,the research on pedestrian vehicle accidents not only focuses on pedestrians,but also on drivers’ risk perception of pedestrian crossing.The details are as follows:(1)Based on the theory of driver’s risk perception and risk prediction,a 19 km long mountain road simulation road is constructed by using the Weidi road aided design system and the vs design software designed by the driving simulation laboratory.The pedestrian crossing is divided into three types: passing through the village section,the roadside vehicles blocking other pedestrians crossing the street,passing through the section with poor sight distance at the intersection,Finally,12 typical pedestrian crossing experimental scenes were built;According to the questionnaire,there are 13 safe drivers,13 general drivers and 13 dangerous drivers.(2)Through the subjective evaluation questionnaire of 39 drivers,the influence of risk perception on drivers’ psychology and driving habits was investigated,and the reliability and validity of the questionnaire were analyzed.The results showed that the reliability coefficient was 0.720≥0.7,and the validity kmo value was 0.717≥0.6,indicating that the designed questionnaire was effective;The quantitative value of risk perception in each scenario is studied,and the risk perception degree of each driver in12 scenarios is obtained;After that,the stepwise regression model was used to extract36 influencing factors with high risk perception.(3)This paper studies the relationship between driver’s driving behavior data,eye movement and psychological indicators and risk prediction ability,analyzes the characteristics of driver’s risk prediction by combining risk perception and decisionmaking methods,and constructs the evaluation and prediction model of driver’s risk prediction ability through pca-ahp theory.Finally,the average score of risk prediction of all kinds of drivers was obtained by weighted calculation.The weighted scores of risk prediction ability of safe,general and dangerous drivers were 7.7515/10,6.0524/10 and 5.0274/10 respectively.The support vector machine algorithm in machine learning is applied to predict the evaluation model.The accuracy of the model is 86.5%,which proves that the model is effective.In this paper,the risk perception of different types of drivers under the mountain road pedestrian crossing is taken as the research object,and the behavior changes of different types of drivers under the specific crossing conflict are studied.Based on the investigation and analysis of drivers’ eye movement characteristics,physiological characteristics,driving behavior characteristics and traffic scene characteristics,this paper puts forward a risk prediction ability evaluation and prediction model suitable for Chinese drivers under pedestrian crossing,and studies its risk prediction ability to deal with pedestrian crossing,which is an important means to improve the driving safety of Chinese drivers,It provides a reference for the construction of automatic driving model under the condition of vehicle road coordinated pedestrian crossing.