Highway Traffic Safety Evaluation Based on TOPSIS-entropy Weight Method and Aerial Data

The research and assessment of traffic accident prevention risks primarily rely on analyzing historical accident data to conduct posture statistics and predict trends. Still, the occurrence of traffic accidents has a strong randomness and low probability, it is still highly unpredictable. Getting th...

Full description

Saved in:
Bibliographic Details
Published in2024 12th International Conference on Traffic and Logistic Engineering (ICTLE) pp. 283 - 288
Main Authors Chen, Chao, Xu, Chuyan, Zhang, Zhengping, Zhi, Ye, Wang, Ziyu, Lyu, Zhouhang, Zhang, Yuxin
Format Conference Proceeding
LanguageEnglish
Published IEEE 23.08.2024
Subjects
Accidents
aerial data
Autonomous aerial vehicles
Evaluation models
Indexes
Road transportation
Safety
safety evaluation
Statistical analysis
Surveys
TOPSIS-entropy weight method
traffic conflict
Trajectory
Vehicle dynamics
Online AccessGet full text

Cover

More Information
Summary:The research and assessment of traffic accident prevention risks primarily rely on analyzing historical accident data to conduct posture statistics and predict trends. Still, the occurrence of traffic accidents has a strong randomness and low probability, it is still highly unpredictable. Getting the accident data to meet the statistical analysis requirements takes a long time. This paper proposes a safety evaluation method for highway traffic, the evaluation index includes average speed, traffic flow, traffic composition, number of lane changes, Number of conflicts based on the TTC, Time Exposed Time to-collision and Time Exposed Time Headway. UAV (Unmanned Aerial Vehicle) aerial survey data is used in this paper, from which the driving track of vehicles on a section of highway, and traffic conflict indicators and their threshold can be extracted. What's more, TOPSIS-entropy weight method is proposed to establish the safety evaluation model of the highway dynamic scenario. Finally, the practical value of the evaluation model is verified through different scenario examples.
DOI:10.1109/ICTLE62418.2024.10703903