Construction of Collision-Type Prediction Models Based on Pre-Crash Data for Advanced Driver Assistance Systems
Construction of Collision-Type Prediction Models Based on Pre-Crash Data for Advanced Driver Assistance Systems
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.)：￥1,100 Members (tax incl.)：￥880
- Publication code
- 20225266
- Paper/Info type
- Proceedings (Spring)
No.60-22
- Pages
- 1-2(Total 2 p)
- Date of publication
- May 2022
- Publisher
- JSAE
- Language
- English
- Event
- 2022 JSAE Annual Congress (Spring)
Detailed Information
| Author(E) | 1) Junhao Wei, 2) Yusuke Miyazaki, 3) Kouji Kitamura, 4) Fusako Sato |
|---|---|
| Affiliation(E) | 1) Tokyo Institute of Technology, 2) Tokyo Institute of Technology, 3) AIST, 4) JARI |
| Abstract(J) | この発表ではアメリカのNASS-CDS及びCISSデータベースを用いて、衝突前情報に基づく衝突形態予測モデルの構築について展開する。 ロジスティク回帰、SVMなどの伝統的なモデルと決定木をベースとするアンサンブル学習手段LGBMが行われ、それぞれの結果の比較することで、特徴とモデルの選択を行い、推定精度のあるモデルの構築しました。 Translation |
| Abstract(E) | In this study, a crash configuration prediction model based on pre-crash information was developed using the accident databases NASS-CDS and CISS in US. Decision tree-based ensemble learning LGBM was performed together with traditional methods such as logistic regression and SVM. Comparison of the results obtained from each model, while investigating the features that increase the accuracy, showed that the prediction model built with LGBM had the highest accuracy. |
