Injury Severity Prediction Based on Select Vehicle Category of Real-World Accidents Data
米国事故データを用いた車両カテゴリーを選定した重症度予測
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.)：￥1,100 Members (tax incl.)：￥880
- Publication code
- 20225047
- Paper/Info type
- Proceedings (Spring)
No.11-22
- Pages
- 1-6(Total 6 p)
- Date of publication
- May 2022
- Publisher
- JSAE
- Language
- Japanese
- Event
- 2022 JSAE Annual Congress (Spring)
Detailed Information
| Author(J) | 1) 江島 晋, 2) 後藤 司, 3) Peng Zhang, 4) Kristen Cunningham, 5) Stewart Wang |
|---|---|
| Author(E) | 1) Susumu Ejima, 2) Tsukasa Goto, 3) Peng Zhang, 4) Kristen Cunningham, 5) Stewart Wang |
| Affiliation(J) | 1) SUBARU, 2) SUBARU, 3) University of Michigan, 4) University of Michigan, 5) University of Michigan |
| Affiliation(E) | 1) SUBARU, 2) SUBARU, 3) University of Michigan, 4) University of Michigan, 5) University of Michigan |
| Abstract(J) | NASS-CDSから衝突時の乗員傷害を予測するアルゴリズムを開発した。対象車両はSUBARU車が属する車両カテゴリーを選び，重傷(ISS 15+)を負う確率を予測した．モデルの感度と特異度は45.1％と96.6％であり，有意な因子は速度変化，ベルト有無，年齢であった．また，助手席乗員の影響が大きく，側突の場合に顕著にみられることがわかった Translation |
| Abstract(E) | An Injury Severity Prediction algorithm was developed using logistic regression to predict probability of sustaining a severe injury. National Automotive Sampling System Crashworthiness Data System data (1999-2015) was filtered for new case selection criteria, which are based on vehicle body type, to match Subaru’s fleet with 2000 model year or over. This proposed algorithm uses crash and driver factors as well as the presence of a right-sided passenger. Model sensitivity and specificity were 45.1% and 96.6%, respectively. Right-front passenger presence is a significant injury risk modifier especially for side impacts (nearside/far-side). |
