基于自组织K-means的城市道路VRU事故场景复杂度评价

打开文本图片集

关键词：弱势道路使用者（VRU）；智能汽车；典型场景；自组织K-means聚类分析中图分类号：U467 文献标识码：A DOI：10.3969/j.issn.1674-8484.2025.03.004

Abstract：Inorder to address the requirements of high-risk testing environments forvalidating inteligent vehicle collsion avoidance systems，while simultaneously to enrich the content and methods for evaluating autonomous driving scenarios involving vulnerable road users （VRU）.This studycollectedand systematically analyzed trafficaccidentcases inGuilinCityGuangxiProvince，from2016 to2020.A totalof1429vehicle

VRUcollsionaccident data werescreened.Based onaccident investigationexperience，13 risk factors were identified，and10typical vehicle-VRUcollsionscenariosapplicable tourbantraffcconditionsinwere constructed using self-organizing K-meansclusteringanalysis.Anevaluation modelforthe complexityof VRU scenarios was established utilizing information entropy theory.The stateof variablesand theweightof each dimension were determined through a combination of logistic regression modelsand back propagation （BP） neural networks，and thecomplexityof various scenarios wascalculated.Aditionally，the Gaussianmixture model was employed tocluster thecomplexity levels，resulting in four distinct scenecomplexitycategories.The results show that on roads with a speed limit of 30km/h ，the nighttime side collsion betweena straight-moving carandanelectric bicycle crossing the road outsideapedestriancrossing areais the most complex scenario. Thefindings inthis study providean experimental scenarioreflectiveof urbanroadcharacteristics in for intelligent vehiclesafety testingandoffera basis for the formulationofexternalVRUcollsionavoidance strategies and decision-making.

Key words： vulnerable road users （VRU）; intelligent vehicles; typical hazardous scenarios; self-organizing; k-means clustering analysis

随着自动驾驶技术的迅速发展，智能汽车正逐渐成为交通系统的重要组成部分。（剩余15357字）