Risk characteristics analysis of road segments: Considering multiple scales and temporal stages

Area of Science:

• Traffic safety analysis
• Road safety engineering
• Transportation systems analysis

Background:

• Traditional traffic conflict analysis often uses discrete models.
• Understanding the dynamic, multi-scale nature of traffic conflicts is crucial for safety.

Purpose of the Study:

• To propose a continuous multi-scale method for analyzing traffic conflict risk characteristics.
• To investigate the spatiotemporal evolution of traffic risk.

Main Methods:

• Defined three scales of traffic entities (vehicle pair, cluster, group) based on interaction range.
• Developed risk expression models for each scale and a dynamic sequential structure for temporal processes.
• Employed Spearman correlation, Friedman test, and multinomial Logistic regression to analyze risk patterns and influencing factors.

Main Results:

• Traffic risk levels exhibit dynamic variations, not strictly monotonic changes.
• Significant differences in risk characteristics were observed across spatial scales and temporal stages of conflicts.
• Unimodal low-risk and high-risk patterns were dominant, with mean speed being a critical precursor.

Conclusions:

• The study provides a multi-dimensional analysis of traffic conflict development across scales and time.
• Differences in spatiotemporal risk evolution among traffic entities were identified.
• The findings offer insights for enhanced road traffic safety management.