歩行者および自転車乗員の軽傷・重傷・死亡別 の新しい傷害予測アルゴリズムの開発

In this study, two injury prediction algorithms for an Advanced Automatic Collision Notification (AACN) system were developed for the road safety benefit of pedestrians and cyclists. The algorithm development data consisted of all car-to-pedestrian and car-to-bicycle accidents that occurred throughout Japan (2014-2019), using macro data from the Institute for Traffic Accident Research and Data Analysis (ITARDA). The injury prediction algorithm for pedestrians was developed by considering the main pedestrian injury risk categories including vehicle travel speed, vehicle type, type of road, pedestrian behavior, pedestrian age and in addition, natural lighting conditions and the main injured body area of the pedestrian. For cyclists, vehicle travel speed, vehicle type and age of cyclist were considered as well as additional injury risk factors which included cyclist behavior, impact direction, helmet usage, natural lighting conditions and the main injured body area of the cyclist. The predicted injury values (objective variables), were used in an ordinal logistic regression for minor injuries, serious injuries and fatalities. In this study, the injury prediction algorithm for pedestrians is referred to as Version 2023P and for cyclists as Version 2023C. In emergency medical care, treatment priorities are determined based on the severity and urgency of the injured patient, known as triage. The priority level differs greatly between a patient with a serious injury requiring urgent treatment and a patient with no signs of life. The injury prediction algorithm developed in this study can represent risk curves for minor injuries, serious injuries and fatalities, and can be applied for triage determination at the time of an accident by incorporating it into an AACN system. [ABSTRACT FROM AUTHOR]