Loading...

Safety Evaluation and Accident Number Estimation at Unsignalized Urban Intersections using Traffic Conflict Indicators

Shukuhikia, Mohammad Amin | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57986 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Nasiri, Habibollah
  7. Abstract:
  8. Serious concerns in the field of road safety, with over 1.35 million fatalities and 50 million injuries annually worldwide 90% of which occur in developing countries such as Iran highlight the importance of research in this area. Also, limitations in accident data have driven studies focusing on preventive methods such as traffic conflict analysis. Furthermore, the application of conflict indicators with various natures and their utilization in estimating the severity and frequency of accidents, particularly at unsignalized intersections due to the nature of maneuvers, present promising opportunities for deeper investigation. The aim of this study is to evaluate factors influencing the severity of traffic conflicts and estimate the number of accidents at urban unsignalized intersections using traffic conflict indicators. In this research, video recordings were conducted at identified intersections in city of Tabriz. These videos were processed using T-Analyst software to identify various variables, including traffic conflict indicators related to the time proximity to collision and movement characteristics. A total of 606 conflicts and the related parameters were identified, leading to the development of two models. A multinomial logistic regression model was used to identify factors affecting conflict severity. In this regard, PET, Delta V, and the mass ratio of vehicles showed significant relationships with conflict severity. Additionally, using the Peaks Over Threshold (POT) approach from extreme value theory, annual accident frequencies at the two studied intersections were estimated using three indicators: TTC, PET, and T2. These estimations were compared with recorded accident data. The results demonstrated the high accuracy of the POT approach in predicting the number of accidents at urban unsignalized intersections. Among the indicators, T2, PET, and TTC exhibited the highest prediction accuracy, respectively. Finally, safety measures were proposed to reduce the number and consequences of accidents
  9. Keywords:
  10. Traffic Safety ; Unsignalized Intersection ; Extreme Value Theory ; Peak Over Threshold ; Conflict Severity ; Multinomial Logistic Regression Model ; Traffic Conflict

 Digital Object List

 Bookmark

No TOC