Research On Abnormal Driving Behavior Of Mountain City Road Under Typical Weather Conditions

With the deterioration of weather conditions,the originally changeable driving environment will become more complex,and the conflict between vehicles will become more prominent,and the traffic system operation is more chaotic,increasing the risk of road traffic safety.The massive data resources accumulated by the development of Internet of Vehicles technology also lead to many new ideas for the study of traffic safety issues.Based on GPS data and driving behavior data in the Internet of Vehicles,this paper analyzed the influence of weather conditions,the combination of weather conditions and road factors,and the combination of weather conditions and road factors on the abnormal driving behavior rate of vehicles step by step.Finally,the significant factors affecting abnormal driving behaviors were determined,and the abnormal driving behavior rate model under typical weather was established.Combined with the road safety risk assessment method,the risk state of road traffic safety was identified,so as to provide data support and decision-making basis for the traffic safety management and technical research department.Specific research content includes the following three aspects:(1)Sunny day,cloudy day,high temperature day,rainy day and foggy day were taken as the main factors in this study.Firstly,according to the unique road conditions of mountainous cities,the typical study sections were selected.Then,under different weather conditions,the statistical changes and spatial distribution of abnormal driving behaviors in each road section were tested to analyze the change rules of abnormal driving behaviors in different weather conditions.Finally,the typical weather conditions that have significant influence on abnormal driving behavior were determined.The research showed that the change trend of abnormal driving behavior rate was not obvious in sunny days,cloudy days and hot days,indicating that drivers would not be influenced to change their driving behavior too much in hot days and cloudy days.The influence of fog on abnormal driving behavior was mainly concentrated on expressways.Drivers would be more cautious when driving due to the excessively fast speed and single reference.The influence of rainy weather on abnormal driving behavior showed obvious regularity.The study found that in rainy days,most groups would choose private cars to travel.Although the number of abnormal driving behaviors increased,the rate of abnormal driving behaviors decreased,and with the increase of the degree of rainy days,it showed a gradually declining trend,indicating that most drivers would maintain a more cautious driving state in rainy days.Therefore,the use of Internet of Vehicles data had significant significance for mining abnormal driving behavior rules and road traffic safety risks on urban roads in rainy days.(2)Taking rainy day as a typical weather factor,the influence of rainy day on four abnormal driving behaviors was studied.In this study,a correlation study was conducted between rainy day conditions and road conditions,and detailed analysis was carried out from multiple levels such as single factor,double factor and multi-factor.Finally,according to the results of association rules,the significant factors affecting abnormal driving behaviors in rainy day were determined.Based on the rainy day to study the influence of abnormal driving behavior,the results showed that the rain can lead to behavior increase,snap acceleration,deceleration,lack of speed,less effect on the number of turn behavior,due to the rainy day more travelers to choose private cars,leading to snap acceleration rate,and decrease rate,sharp turn rate,super rate decreased.The results showed that the sharp acceleration and deceleration behavior are related to a variety of road factors,not only affected by the two factors,the steep slope had a significant impact on the sharp deceleration behavior,and increases with the enhancement of the weather degree.The sharp turn behavior was not only affected by the curve,but also needs the joint action of other road factors.Steep slope had great influence on abnormal driving behavior.The result of multi-factor comprehensive influence showed that the abnormal driving behavior was the result of multi-factor combined effect,and its influence result was greater than the effect of single factor and double factor.The rapid acceleration behavior was affected by various road factors such as slope,camber and opening,and the frequency of rapid acceleration increases with the increase of rainfall intensity.The abrupt deceleration behavior was affected by various road factors such as slope,camber and opening,and the frequency of abrupt deceleration increases with the increase of rainfall intensity,and the influence degree was greater than that of abrupt deceleration behavior.The abrupt turn behavior was mainly affected by slope,camber,opening and other road factors,but the influence of rainfall intensity was not significant.The overspeed behavior was affected by many road factors,such as slope,camber and information control opening,and the frequency of overspeed behavior decreased with the increase of rainfall intensity.(3)Slope,camber and the number of openings were used as quantitative variables,while information control openings and rainfall intensity were used as dummy variables.The Poisson regression model and negative binomial regression model were tested respectively.Finally,four abnormal driving behavior rate models of rapid acceleration,rapid deceleration,sharp turning and overspeed were established.Combined with the road safety evaluation method and the safety risk classification method,the road traffic safety risks under rainy days were identified.The research showed that the negative binomial regression model was better than the Poisson regression model to establish the abnormal driving behavior rate model in rainy days.Finally,Chongqing Jingwei Avenue,Yulu Avenue and Changjianger Road were selected to be divided into 23 road sections.The rapid acceleration rate,rapid deceleration rate,sharp turn rate and overspeed of each road section were extracted,and the regression model of abnormal driving behavior rate established and the road traffic safety identification method were verified by examples.The results showed that the abnormal driving behavior rate can ensure the accuracy of the model within the error range.The classification threshold was 0.040 and the accuracy reached 91.3% under light rain conditions,the classification threshold was 0.037 and the accuracy reached 86.96%under moderate rain conditions,and the classification threshold was 0.030 and the accuracy reached 95.65% under heavy rain conditions.