Comparison of Two Models Evaluating Automobile Safety Features

Abstract

Two existing models of automobile crashes are extended and compared to improve the evaluation of the performance of new, vehicle-based technologies for crash avoidance. The crashes model is based on knowledge of the frequencies and severities of common crash scenarios. The fault tree model is based on knowledge of the frequencies of and relationships among accident causation events. New technology countermeasures, such as drowsy-driver warning and center-mounted brake lights, can be studied individually and in combination using either model. The effectiveness of the countermeasure technologies is measured as the reduction in the frequency and cumulative harm of crashes. For some technologies, it is more straightforward to estimate the effect of technology to mitigate a well-defined accident scenario, as in the crashes model. For other prospective technologies, it also is possible to measure the influence of technology on a well-defined initiating event, as in the fault tree model.