Research On Recognition Of Driving Style In Stable Car-following State

With the continuous development of society and the continuous improvement of road infrastructure,the rapid growth of car ownership has brought great challenges to the smoothness of urban traffic and the degree of road traffic safety.According to my country’s road traffic safety statistics annual report data,nearly 85% of traffic accidents in my country are related to drivers,and most of them are rear-end collisions,while rear-end collisions on expressways are characterized by high harm and serious economic losses.Therefore,based on the stable following condition of Expressway driving,starting from the driving style of drivers,this paper explores the vehicle motion characteristics of different styles of drivers in Expressway driving,and establishes the driving stability evaluation system of different styles of drivers under the stable following condition.Based on the domestic and foreign researches on stable following state and driver’s style,this article first clarifies the purpose of the actual vehicle test.Combining the actual traffic conditions of my country’s expressways and the characteristics of Chinese drivers,select suitable test sections,design test plans,and then recruit drivers of different ages,driving ages,and genders for real-vehicle tests.Collect key indicators such as velocity,following distance,Acceleration and yaw rate,and screen out stable following events based on theoretical research,and then analyze the driving behavior index rules of stable following events.Furthermore,K-means clustering is used to classify driving styles into conservative,general and aggressive driving styles,and then statistical analysis and correlation analysis are performed on driving behavior indicators of different styles of drivers to explore the differences in these indicators of different styles of drivers.Finally,the entropy method is used to establish a driving stability evaluation system for drivers.This article mainly draws the following conclusions:(1)Correlation analysis and coupling analysis of various indicators under stable following car conditions are carried out.The results show that there are significant differences in accelerator pedal opening,following distance and yaw rate under different vehicle speed intervals,and significant differences in vehicle speed,accelerator pedal opening and time following in different following distances;lateral movement indicators such as standard deviation of yaw rate,standard deviation of lateral acceleration and standard deviation of lateral offset,etc.,show similar laws in different vehicle speeds,distances and time intervals.(2)Using the three indicators of following time,following distance and accelerator pedal opening to perform K-means Clustering on the driver’s style,and using the Decision Tree algorithm to test the clustering results,the results show that the overall accuracy rate is 99.7%;a statistical analysis of the driving behavior indicators of drivers of all styles reveals that under stable following conditions,the more aggressive the driving style,the more aggressive the car’s velocity,acceleration,and angular velocity when the accelerator pedal is on.The greater the degree,the smaller the following distance and the following time between the following vehicle;the correlation analysis shows that different driving styles and different speed ranges have a greater impact on indicators such as vehicle velocity,following distance,lateral/longitudinal acceleration,and yaw rate when the driver is driving;then analyzing the changes in driver’s style under long-term working conditions,it is found that conservative drivers will be more conservative over time,aggressive drivers are more aggressive,and ordinary drivers have stable changes in style,but tend to be conservative.(3)Based on the previous research,the standard deviation of the four indicators of lateral acceleration,lateral offset,longitudinal acceleration,and following distance are determined as indicators to establish a stability evaluation system for drivers of different styles under stable following conditions,through Entropy Method assigns weights to four indicators,and then scores the driving stability of 20 drivers.The evaluation result obtained by this method agrees with the driving style of 90% overall,and the driving stability of conservative drivers and ordinary drivers crosses;male drivers have less conservative styles,driving stability and driving The style is strongly correlated.Female drivers have a relatively large proportion of conservative styles,and the driving stability is strongly correlated with driving proficiency.