Drivers passing cyclists: How does sight distance affect safety? Results from a naturalistic study.

Area of Science:

• Road safety research
• Traffic engineering
• Human factors in transportation

Background:

• Cycling offers numerous benefits but poses risks due to shared roads with motorized traffic.
• Overtaking cyclists is a critical maneuver for drivers, influenced by lateral clearance, speed, and environmental factors like sight distance and oncoming vehicles.
• Driver behavior during overtaking can negatively impact cyclist safety and their willingness to cycle.

Purpose of the Study:

• To investigate the impact of visibility, specifically sight distance, on driver behavior when overtaking cyclists.
• To analyze how factors like sight distance and oncoming traffic influence drivers' lateral clearance and speed during overtaking maneuvers.
• To provide data for infrastructure design, policymaking, and the development of active-safety systems.

Main Methods:

• Collected naturalistic driving data of drivers overtaking cyclists on a rural Swedish road using four roadside traffic sensors.
• Employed a Bayesian multivariate approach to model lateral clearance and speed at the moment of passing.
• The model was fitted on 81 overtaking maneuvers, analyzing the influence of sight distance and oncoming traffic.

Main Results:

• Reduced sight distance was correlated with drivers reducing their lateral clearance when overtaking cyclists.
• Driver speed was also reduced under reduced sight distance, though less pronounced than clearance changes.
• The presence of oncoming traffic had a significant effect, similar to but stronger than reduced sight distance, on overtaking behavior. Some critical low clearances were observed.

Conclusions:

• Cyclist safety is compromised by drivers overtaking under conditions of reduced visibility or in close proximity to oncoming traffic.
• Infrastructure and policy decisions should consider prohibiting overtaking in areas with limited sight distance or oncoming traffic.
• The developed model and collected data can inform active-safety systems, automated driving development, and driver model evaluation.