Research On UDS-based Driving Status Evaluation System

With the development of intelligent automobiles in the Industry 4.0,the ownership of vehicles has grown rapidly.At the same time,the safety issue of traffic has gradually become the focus of global attention.The main factors affecting current traffic accidents can be summarized into three categories:driving environment,driving habits and vehicle breakdowns,which traffic accidents caused by bad driving habits account for more than 70%.The level of driving status is directly related to the safety of road traffic.It is necessary to study the status classification system for traffic conditions.Due to inconsistent collection methods,analysis methods,and modeling algorithms for driving data.The research methods for driving behavior evaluation have not yet reached a unified standard at home and abroad.In order to fundamentally clarify whether the traffic safety problem is caused by personal behavioral factors or due to the vehicle’s own failure,this thesis synthesizes the changing laws of multi-dimensional reference indicators during the driving,extracts various index characteristics,and proposes a three-dimensional driving habit decision method based on safety,economy,and comfort.The driving status are classified into three categories:good,bad and dangerous.The research process of this thesis includes the following four stages.The first stage analyzes the relationship between driver’s habits and perception of the environment in the driver’s environment perception-brain decision-operational execution cycle.Based on the Analytic Hierarchy Process(AHP),a multi-dimensional decision model for driving behavior is established,and a three-dimensional evaluation system for safety,economy,and comfort was proposed.In the second stage,the vehicle information acquisition equipment,the prep-rocessing data,and the extraction features are designed.The driving data is derived from the real-time traffic information read by the vehicle information acquisition equipment under real road conditions.In the third stage,the AHP-SVM traveling status assessment model is built using the MATLAB tool platform to classify the driving status into two categories:good and bad.The fourth stage analyzes the dangerous driving state evaluation standard in driving anomaly,and establishes a dangerous state assessment model based on a Support Vector Machine(SVM)to identify dangerous conditions in driving conditions.The test results show that the AHP-SVM algorithm model has the best performance using Gaussian kernel function(RBF).The system divides driving habits into three categories:good driving condition,bad driving condition and dangerous driving condition.The classification accuracy rates from comfort,economy,and safety are 99.4918%,97.2897%and 84.5285%,the overall accuracy is 88.6872%.And the accuracy rate of identifying danger is 90.6268%.It proved the practicality of the system and achieved the purpose of reminding the driver to improve the bad and dangerous driving status from the perspective of his own driving behavior and vehicle health status,thereby improving the traffic safety problem.