Citation

Tominaga S, Nishimoto T, Motomura T, Matsumoto H, Lubbe N, Kiuchi T. Trans. Soc. Automot. Eng. Jpn. 2015; 46(5): 925-930.

Copyright

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

DOI

10.11351/jsaeronbun.46.925

PMID

unavailable

Abstract

In this paper, injury prediction algorithm for Advanced Automatic Collision Notification (AACN) based on Japanese traffic accident situation was developed. AACN algorithm was applied to Japanese large accident database by using a logistic regression modeling technique. Risk factors associated with driver's risk of severe injury were Delta V, crash direction, seat belt use, multiple impact, occupant's age, vehicle class(Passenger car, Kei car). Cut off value of severe injury risk of AACN algorithm was clarified by using ROC analysis. In addition, In-depth accident data base was applied to AACN algorithm to compare a relation between algorithm output and actual injury results.

Language: ja

Keywords

Automatic collision notification; In-depth accident investigation data; Injury prediction; Logistic regression; Safety