A Study Of Evaluating Of Driving Capability Based On Driving Simulator And Its Applications

In the transportation system which is composed of human,vehicle and road,human is the most critical part and it is also the most important factor leading to traffic accident.Therefore,the evaluating of driver’s driving capacity can be used to analyze the human cause mechanism of traffic accident and provide a theoretical basic for the formulation of corresponding countermeasures.Hence,this study will collect vehicle state data based on driving simulation test,research the evaluating methods of the drivers’ fundamental driving capacity and the drivers’ dynamic driving capacity,and explore the applications of driving capacity evaluating.Firstly,a virtual simulation model was built based on UC-Win/Road and the driving simulation tests was carried out.Three driving scenarios were designed in the virtual simulation model: the longitudinal driving capacity test scenario,the lateral driving capacity test scenario and a dangerous event test scenario which contains 14 dangerous events of 6 categories.In additional,a subjective questionnaire on driving ability was designed which includes five aspects: basic information,security awareness,radicalization,irritability and personality characteristics,so as to compare and analyze with the results of driving simulation test.Secondly,the Fundamental Driving Capacity evaluating method is established.The driving capability is estimated from three typical driving behaviors:longitudinal driving,lateral driving and lane changing.On this basis,the characteristic parameters of the fundamental driving capacity are determined,the subjective weight of the parameters is established by the Analytic Hierarchy Process(AHP),and the objective weight of the parameters is established by the Entropy Weight Method(EWM),and then the comprehensive weight is obtained.Finally,the fundamental driving ability of drivers is evaluated based on the TOPSIS method.Thirdly,the Dynamic Driving Capacity evaluating method is established.Taking into account the differences between individual drivers,a safe driving behavior prediction model based on a LSTM neural network is established by using safe driving data.The driving behavior deviation is defined,and the adjustment amount of driving behavior deviation is given.Based on the driving behavior deviation adjustment amount and the vehicle safety state indicator,a dynamic driving capability evaluation model is established.Lastly,the comparative analysis of fundamental driving capacity and driving behavior questionnaire,and the comparative analysis of fundamental driving capacity and dynamic driving capacity are carried out.On this basis,combined with the fundamental driving capacity and the dynamic driving capacity,the risk warning strategy which can be applied to the assisted driving system is developed,and the effectiveness and accuracy of the risk warning strategy are analyzed.