Citation

Nishimoto T, Mukaigawa K, Tominaga S, Kiuch T. Trans. Soc. Automot. Eng. Jpn. 2015; 46(6): 1123-1129.

Copyright

(Copyright © 2015, Society of Automotive Engineers of Japan)

DOI

10.11351/jsaeronbun.46.1123

PMID

unavailable

Abstract

The purpose of this study was to establish an injury prediction algorithm for an Advanced Automatic Collision Notification (AACN) system for vulnerable road users such as the pedestrian and bicycle users. The injury prediction algorithm was based on two independent samples of ITARDA macro data, one of which was used for training and the other as for validation. The AACN algorithm was developed using the Japanese large scale accident database by a logistic regression modeling technique. The Risk factors associated with severe injury for pedestrians and cyclists were travel speed, the frontal shape of striking vehicle, pedestrian and cyclist age, type of road and pedestrian and cyclist behavior. Validation of both AACN algorithms were verified using ROC analysis. The results indicate that for a 10% rate of under triage, the threshold values are 8.8% and 2.9% for pedestrians and cyclists respectively.

Language: ja

Keywords

Automatic collision notification; Cyclist; Injury prediction; Macro data; Pedestrian; Risk curve; Safety