Analyzing drivers' hazard recognition: Precursors to single-vehicle collisions

Area of Science:

• Traffic Safety
• Human Factors in Driving
• Crash Avoidance Research

Background:

• Real-world driving data from the SHRP2 Naturalistic Driving Study reveals complex factors influencing crash outcomes.
• Understanding why drivers fail to recognize precipitating events (PEs) is key to improving road safety, especially in single-vehicle crashes where recognition rates are low.

Purpose of the Study:

• To investigate factors correlated with the recognition of precipitating events in single-vehicle crashes.
• To analyze how these factors act as crash precursors and influence crash outcomes.
• To identify specific scenarios where driver recognition is diminished.

Main Methods:

• Utilized a path-analytic framework with marginal effects to examine correlations between variables and PE recognition.
• Employed logit models with fixed and random parameters to account for event and driver heterogeneity.
• Analyzed data from 543 crash or near-crash events.

Main Results:

• Hazard type significantly impacts recognition: less conspicuous hazards (e.g., departing roadway) decreased recognition by 48.29%, while prominent hazards (e.g., loss of control) increased it by 46.71%.
• Driving environment and maneuvers play a role: relaxed conditions like daylight decreased recognition by 16.00%, whereas attention-demanding environments improved it.
• Recognition of PEs reduced crash likelihood by 12.23%.

Conclusions:

• Driver recognition of precipitating events is strongly influenced by hazard prominence, roadway environment, and driving maneuvers.
• Interventions should aim to enhance driver awareness of less conspicuous hazards and in less demanding driving situations.
• Future intelligent transportation systems can be designed to proactively increase driver recognition of potential road hazards.