ARSRPE Conference Paper Database

An Integrated Driver-Vehicle-Environment (I-DVE) model to assess crash risks

Glaser, Sebastien, Rakotonirainy, Andry, Gruyer, Dominique, Nouveliere, Lydie



A wide range of driver and vehicle models have been proposed by traffic psychologists, engineers and traffic simulation researchers to assess crash risks. However, existing approaches are often confined within a single discipline and lack concepts that formally express the complexity of interactions between the driver, vehicle and environment as well as the broader scope and the interdisciplinary nature of the driving behaviour modeling. For example, traffic psychologists have defined a driver performance model as the driver's perceptual and motor skills (capabilities), or what the driver can do. In contrast, a driver behavior model refers to what the driver actually does do while driving (Evans, 1991). A driver behaviour model is determined by an infinite and complex number of factors related to the environment, driver and vehicle but is not explicitly modeled in Evans (1991). Existing driver models lack substantive concepts that express the interactions between the Driver, Vehicle and Environment (DVE). A new Integrated Driver-Vehicle-Environment (I-DVE) model is formally presented as a set of concepts and equations representing interactions between the driver, vehicle and environment with the view to assess crash risks. The I-DVE model features realistic and measurable attributes, which ultimately influence the driving performance and associated crash risks. I-DVE model is validated in a simulation. The simulation uses empirical data related to Time To Collision (TTC), Energy Equivalent Speed (EES), injury severity and driver profile to assess crash risks. This paper (i) reviews existing driver modeling approaches and highlights the need for an integrated approach, (ii) defines a novel model capable of expressing risks associated interaction between the driver, environment and vehicle and (iii) provides directions for further research in driver behaviour modeling.