An Attention-Based Model of Driver Performance in Rear-End Collision Situations

This paper demonstrates the validation of a model of the driver that was developed to help address issues concerning the prevalence of rear-end collisions. The model validated here is the implementation longitudinal control of a vehicle based on Gibson’s Field Theory of driving. It was found that the field theoretic models of the driver could be successfully implemented to examine longitudinal control. It was also determined that the field-theoretic model could accurately predict most of the findings from a more traditional simulator approach to examining collision situations. The field-theoretic model was able to predict difference not only as to the outcome of the situation, but also with regard to the process used by the driver to reach that outcome. This implementation of the field-theoretic model of driving was not perfect however, and there were several areas where the model could not produce results that matched the simulator study. Overall, the field-theoretic approach to modeling was able to produce valid and accurate predictions of driver performance in terms of both process and outcome. This field-theoretic model of the driver will provide a valuable tool for examining a wide range of issues that are not necessarily limited to rear-end collisions.