The Research On Driving Behavior-Speed Control Based On Vehicle Networking Data

The study of driving behavior has become the core content of the national road traffic safety research due to the driver's core position in the whole accident prevention work.But the evaluation of the speed control behavior exists light connotation and heavy concept,it also doesn't form a evaluation index system,analysis content often confined to overspeed behavior combined with limitation of the experimental method,the paper has carried out the following work aiming at this problem.The evaluation system is established according to the characteristics of the collected data from the vehicle system of car networking—Sinotruk intelligent system and the select principle of index.Effective methods of filling the missing value,detecting outliers,deleting parking data can be used to preprocess the 40 target vehicles running data that extracted from the system,the data preprocessed would be standardized in accordance with the selected evaluation index.The standard data carried out factor analysis and cluster analysis with analysis software SPSS,40 driver's speed control behavior is clustered.Two unrelated factor is variable speed driving factor and speeding tendency driving factor could be extracted with the principal component extraction method and the two main factor score of 40 drivers is get,the drivers of each type of driving behavior is classified to inexistence,light,medium,heavy four levels with system clustering according to two big main factor score.Finally,the prevention and control strategy is explored proposed to the bad speed control behavior.Warning system for dangerous driving is put forward according to the clustering results,this paper introduces the requirements and brief warning process that the system should meet;The prevention and control measures are put forward from the driver's education,the truck production standard,road running environment,accident prevention.