Scenario Driven Tests And Evaluation Methods Of Drivers’ Hazard Prediction Ability

With the rapid development of China’s social economy,the level of motorization has been continuously improved,the number of motor vehicles and drivers has continued to increase,and the problem of traffic safety has become increasingly prominent.Driving operation ability,hazard prediction ability,stress response ability,and rule cognition ability jointly affect the safety of drivers.Amongst others,the hazard prediction ability(i.e.,the ability of a driver to perceive and predict the potential danger on the road)reflects a driver’s risk awareness,and the traffic accident data and cause analysis show that it is closely related to driving safety.Therefore,it is necessary to conduct the test and evaluation research on drivers’ hazard prediction ability,so that drivers can better understand their own safe driving level,and consciously respond to possible dangers while driving,so as to avoid the occurrence of traffic accidents.Taking the drivers’ hazard prediction ability as the main research object,this study collects the behavior data of drivers in different traffic hazard scenarios under simulated driving circumstance,builds a scenario driven evaluation model of drivers’ hazard prediction ability using data envelopment analysis,and develops an information management platform for comprehensive evaluation of driver safety based on the visualization software.The specific research contents are as follows:First,in the driving simulation environment,several typical potential dangerous scenarios commonly occurred on urban roads are designed,and the evaluation indicators that can characterize drivers’ hazard prediction ability are selected.Based on the simulated driving experiments,the drivers are tested and trained,and the multi-source driving behavior data including vehicle movement data,eye movement data and pedal use data are collected.The multi-source data are then integrated and preprocessed based on Python,and the behavior indicators in each scenario are extracted.Next,in view of the shortcomings of existing research,a scenario driven approach for drivers’ hazard prediction ability evaluation is proposed,and an evaluation model based on the multi-layer data envelopment analysis is constructed.Using the idea of controlled experiments,the performance difference between the trained experimental group drivers and the untrained control group drivers is compared,and the result verifies the effectiveness of the training method.Taking the evaluation of drivers’ hazard prediction ability as an example,the process for determining the core set of scenarios required for evaluation is proposed.Then,for the evaluation of driving operation ability,stress response ability,and rule cognitive ability,the steps of scenario design,indicator selection,model construction are completed.From a global perspective,the overall safety of a driver is comprehensively evaluated by combining all these abilities,and the evaluation results can be used to generate personalized and targeted improvement plans.Finally,based on the visualization software,a comprehensive evaluation information management platform for driver safety is built,which is convenient for drivers to check their own safe driving records and to discover their weak points of driving behavior.The simulated driving scenarios designed in this paper can be used for driver safety testing.The proposed scenario driven evaluation and training method can provide theoretical basis and practical guidance for driver selection and improvement of driving safety.The developed information management platform for comprehensive evaluation of driver safety can provide technical support for driving behavior monitoring of individual drivers,driving schools,transportation companies and traffic management departments.