欢迎访问《汽车安全与节能学报》，
汽车安全与节能学报

汽车安全与节能学报 ›› 2023, Vol. 14 ›› Issue (5): 536-543.DOI: 10.3969/j.issn.1674-8484.2023.05.002

• 汽车安全 • 上一篇    下一篇

基于CART决策树的车辆与行人碰撞中头部损伤风险预测

韩勇1(), 罗金镕1(), 何勇2, 吴贺3, 林旭洁1, 蔡鸿瑜1   

  1. 1.厦门理工学院 机械与汽车工程学院，厦门 361024，中国
    2.厦门华厦学院 信息与智能机电学院，厦门 361024，中国
    3.厦门大学 航空航天学院，厦门 361102，中国
  • 收稿日期:2023-03-04 修回日期:2023-07-17 出版日期:2023-10-31 发布日期:2023-10-31
  • 作者简介:韩勇（1984—），男（汉），江西，教授。E-mail: yonghan@xmut.edu.cn。
    罗金镕（1998—），女（汉），四川，硕士研究生。E-mail: 1007709496@qq.com。
  • 基金资助:
    国家自然科学基金面上项目(51775466);厦门市自然科学基金面上项目(3502Z20227223);司法部司法鉴定重点实验室开放基金(KF202116)

Prediction of pedestrian head injury in vehicle-pedestrian collisions based on a CART decision tree

HAN Yong1(), LUO Jinrong1(), HE Yong2, WU He3, LIN Xujie1, CAI Hongyu1   

  1. 1. School of Mechanical and Automotive Engineering, Xiamen University of Technology, Xiamen 361024, China
    2. School of Information and Intelligent Mechatronics, Xiamen Huaxia University, Xiamen 361024, China
    3. School of Aeronautics and Astronautics, Xiamen University, Xiamen 361102, China
  • Received:2023-03-04 Revised:2023-07-17 Online:2023-10-31 Published:2023-10-31

RichHTML

5

PDF

91

摘要：

为快速预测汽车碰撞行人头部损伤风险，建立了一种基于多刚体系统动力学仿真方法和分类回归决策树（CART）的预测模型。参考欧洲新车评价规程（Euro-NCAP），开发了具有精细化刚度特征的车辆前部结构多体模型；以行人尺寸、初始车辆速度和行人速度、人车碰撞位置、相对角度为变量，通过全因子设计试验方法，建立了4 500组多体仿真模型；采用CART模型，挖掘变量与动力学响应参数的关联性。结果表明：车辆初始碰撞速度是影响行人头部动力学响应的关键因素；该模型对于碰撞速度和头部损伤准则（HIC15）值的预测精度分别为87.5%和86.8%，平均预测耗时为42.7 ms，两者均具有较高的预测精度和决策能力。该结果可为制定行人头部损伤风险评估实验和损伤防护研究提供理论参考依据。

关键词: 汽车安全, 碰撞事故, 头部损伤准则（HIC）, 车辆前部结构, 决策树预测模型, 头部动力学响应, 分类回归决策树（CART）

Abstract:

A prediction model based on a multi-rigid-body system dynamics simulation method and a Classification and Regression Tree (CART) were established to rapidly predict the risk of head injury in vehicle to pedestrian collision. A multi-body model of the vehicle front structure with refined stiffness characteristics was developed with reference to the European New Car Assessment Program (Euro-NCAP). About 4 500 sets of multi-body simulations were established by the full factor design test method with the pedestrian model, initial vehicle speed and pedestrian speed, pedestrian-vehicle collision position, and relative angle as simulation variables, and the CART model was used to explore the correlation between the variables and the kinetic response parameters. The results show that the initial vehicle speed is a key factor affecting the dynamic response of the pedestrian head. The prediction accuracy of the model for the collision speed and the value of the head injury criterion (HIC15) is 87.5% and 86.8%, respectively, and the average prediction time is 42.7 ms, which have high prediction accuracy and decision-making ability. The results can provide a theoretical reference basis for developing pedestrian head injury risk assessment experiments and injury protection research.

Key words: vehicle safety, collision accident, head injury criterion (HIC), vehicle front structure, decision tree prediction model, head dynamics response, classification and regression tree (CART)

中图分类号: