Classification of Crash Pattern Based on Vehicle Acceleration and Prediction Algorithm for Occupant Injury

Abstract

Injury prediction based on data from event data recorder in automatic collision notification is expected to reduce trauma deaths. Known to affect on injury, the crash pattern is required to be classified to accurately predict injury. We performed crash simulations with a full-vehicle finite element model, and determined typical vehicle acceleration profiles of each crash pattern. A method to classify a crash pattern by comparing vehicle acceleration with the typical profiles was found to be effective. We also performed multi-body simulations with a vehicle interior and a dummy model, and developed injury prediction algorithms of each crash pattern.