Analysis of Traffic Conflicts at Roundabout Entrances and Exits - A Machine Learning Approach for Enhanced Safety

Aug 4, 2026

Upload by Eden

Roundabouts form a key part of the urban road network, so their safety performance directly affects overall traffic safety. This study uses UAV video to extract vehicle trajectories at roundabout entrances and exits, then identifies and analyzes traffic conflicts through the time to collision (TTC) index and vehicle evasive actions. An areal-time safety evaluation framework is developed with machine learning models including RF, SVM, XGBoost, and DT. The approach models the relationship between traffic states and conflicts across varying conditions, and evaluates performance using accuracy and AUC. Results show RF achieves the best prediction performance (accuracy 0.86, AUC 0.88) and traffic flow, speed, density, speed standard deviation, and vehicle type ratio are significantly associated with conflict risk.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Abstract

Introduction

Background and motivation

Roundabout safety evaluation and traffic conflicts

Method overview

UAV trajectory acquisition and conflict identification

Machine learning-based safety evaluation models

Analysis of Traffic Conflicts at Roundabout Entrances and Exits – A Machine Learning Approach for Enhanced Safety Yuzhou DUAN1, Zhipeng LIN2, Yulong WANG3, Qiaowen BAI4 Review Submitted: 6 June 2024 Accepted: 13 Dec 2024 [1](1 duanyz@haut.edu.cn)[ duanyz@haut.edu.cn](1 duanyz@haut.edu.cn), College of Civil Engineering, Henan University of Technology, Zhengzhou, China [2](2 linzhipeng0318@163.com)[ linzhipeng0318@163.com](2 linzhipeng0318@163.com), College of Civil Engineering, Henan University of Technology, Zhengzhou, China [3](3 k16657096@163.com)[ k16657096@163.com](3 k16657096@163.com), College of Civil Engineering, Henan University of Technology, Zhengzhou, China 4 Corresponding author, [victorbty@foxmail.com](victorbty@foxmail.com), College of Design and Engineering, National University of Singapore, Singapore This work is licenced under a Creative Commons Attribution 4.0 International Licence. Publisher: Faculty of Transport and Traffic Sciences, University of Zagreb ABSTRACT As a component of the urban road network, roundabouts play a crucial role in ensuring operational efficiency. The safety performance of roundabouts significantly impacts overall traffic safety, making it necessary to conduct safety analysis and evaluation. This study utilises UAV to capture video of vehicle trajectory at roundabouts, employing the time to collision (TTC) index and vehicle evasive actions to identify and analyse traffic conflicts. Areal-time traffic safety evaluation method has been developed using machine learning algorithms, including random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost) and decision tree (DT) model. This method aims to analyse the relationship between traffic states and conflicts, providing insights into potential safety risksin various traffic conditions. The four machine learning algorithms trained a total of 12 models, with RF demonstrating superior training effectiveness. It achieved high accuracy in predicting traffic conflict areas at the entrances and exits of a roundabout, with a prediction accuracy of 0.86 and an AUC (area under the receiver operating characteristic curve) of 0.88 . In addition, this paper further explores the relationship between traffic conflict and state. The results show that traffic flow, speed, density, speed standard deviation and vehicle type ratio have a significant relationship to traffic conflict. This research provides valuable insights for transportation authorities to understand the nature of traffic conflicts at roundabouts, enabling them to implement appropriate early warning systems and management strategies. KEYWORDS roundabouts; machine learning algorithms; traffic states; traffic conflicts. 1. INTRODUCTION All over the world, traffic safety has always been a major concern. In China, according to the National Bureau of Statistics of China in 2022, there were 256,409 road traffic accidents, resulting in 60,676 fatalities and 263,621 injuries, with direct property losses amounting to one billion two hundred million CNY. These statistics highlight the substantial threats to residents ’ lives and property posed by traffic accidents. Urban intersections, accounting for approximately 29.7% of road traffic accidents, pose significant safety challenges. The severity of safety issues at these intersections is particularly notable [1] . Based on the above background, research aimed at reducing and preventing traffic accidents at urban intersections has become a prominent topic within the field of transportation [2-6] . A roundabout is an at-grade intersection with a circular or elliptical central island in the centre. The accident rate of roundabouts is often lower than that of other types of at-grade intersections [7] . Accident records serve as the basic data for evaluating roundabout safety, enabling direct analysis through accident probability. Existing methods for urban roundabout safety analysis and eval

FAQs

How does the study identify traffic conflicts at roundabouts?

It extracts vehicle trajectories from UAV-captured video and uses the time to collision (TTC) index together with vehicle evasive actions to identify and analyze conflicts.

Which machine learning models are used for the safety evaluation framework?

The framework trains random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost), and decision tree (DT) models.

What traffic state variables show a significant relationship with traffic conflicts?

Traffic flow, speed, density, speed standard deviation, and vehicle type ratio have significant relationships with traffic conflict outcomes.

Aug 4, 2026

Upload by Eden