基于人体落地机制预测的人地碰撞损伤防护方法  

Pedestrian-ground Collision Injury Protection Method Based on Pedestrian Landing Mechanism Prediction

作　　者：邹铁方[1] 罗鹏琛 彭培能 胡林[1] 王丹琦 ZOU Tie-fang;LUO Peng-chen;PENG Pei-neng;HU Lin;WANG Dan-qi(School of Automotive and Mechanical Engineering,Changsha University of Science&Technology,Changsha 410114,Hunan,China)

机构地区：[1]长沙理工大学汽车与机械工程学院,湖南长沙410114

出　　处：《中国公路学报》2025年第3期165-176,共12页China Journal of Highway and Transport

基　　金：国家杰出青年科学基金项目(52325211);湖南省教育厅优秀青年项目(23B1111);湖南省普通高等学校科技创新团队项目(2024RC1029)。

摘　　要：车人事故中地面所致行人损伤不可忽视,为降低人车碰撞中地面所致人体损伤及防护风险,提出一种基于人体落地机制预测的人地碰撞损伤防护方法。先基于1300组人车碰撞事故数据建立人体落地机制预测模型,再依据预测结果建立车辆制动规则,最后再依据规则控制车辆运动。720次仿真结果显示,所提方法能准确预测87.5%的人体落地机制,其中机制M1、M2、M3、M4、M5的预测准确率分别为75.29%、95.75%、91.3%、94.55%、100%,且可降低地面所致WIC、HIC分别达61.7%、37.5%,并将方法使用风险降至8.33%。进一步分析具体落地机制预测不准的原因,并据此更新车辆制动规则,改进后的方法能大幅提升M1机制的预测准确率,预测准确率达92.94%,且可将使用风险降至5.83%,地面所致WIC、HIC的降幅则分别提升至70.3%、43.6%。研究成果将为人地碰撞损伤预测提供新思路,为智能车行人保护提供低风险新方法。Pedestrian injuries caused by the ground in pedestrian-vehicle accidents cannot be ignored.Hence,a method of pedestrian injury protection based on pedestrian landing mechanism prediction is proposed to reduce the risk of pedestrian injuries caused by the ground in pedestrian-vehicle collisions.In this study,a pedestrian landing mechanism prediction model was constructed based on 1300 groups of pedestrian-vehicle collision accident data.Vehicle braking rules were established according to the prediction model,and vehicle motion was controlled according to the established rules.The 720 simulation results show that the proposed method can accurately predict 87.5%of the pedestrian landing mechanisms,and the prediction accuracies of M1,M2,M3,M4,and M5 are 75.29%,95.75%,91.3%,94.55%,and 100%,respectively.The ground-related WIC and HIC were reduced to 61.7%and 37.5%,respectively,and the risk was reduced by 8.33%.Further analyses were conducted to determine the reasons for inaccurate predictions of specific landing mechanisms,and the vehicle braking rules were updated based on these results.The improved method can significantly improve the prediction accuracy of M1:its prediction accuracy is 92.94%,and its use risk is reduced to 5.83%.The WIC and HIC caused by the ground decreased to 70.3%and 43.6%,respectively.Thus,the results of this research provide new ideas for predicting pedestrian ground collision injuries and low-risk methods for pedestrian protection against intelligent vehicles.

关 键 词：汽车工程 人车碰撞 人体落地机制 神经网络 人地碰撞损伤 行人损伤防护 

分 类 号：U467.14[机械工程—车辆工程]