Accident prediction model for railway-highway interfaces

Area of Science:

• Transportation Safety
• Traffic Engineering
• Railway Engineering

Background:

• Limited research exists on factors contributing to railway-highway crossing accidents.
• Highway-rail grade crossings account for a significant portion of railway accidents and fatalities in Korea.

Purpose of the Study:

• To identify and analyze factors associated with vehicle accidents at highway-rail grade crossings.
• To compare accident models and safety effects of crossing elements between the US and Korea.

Main Methods:

• Utilized various statistical models, including the gamma probability model, to analyze accident data.
• Examined relationships between crossing accidents and features such as traffic volume, train volume, proximity to commercial areas, train detector distance, and warning signal timing.

Main Results:

• Accident frequency increases with higher total traffic volume and average daily train volumes.
• Proximity to commercial areas, train detector distance, and warning signal activation time are linked to increased accident numbers.
• The gamma probability model effectively addressed underdispersion in the data.

Conclusions:

• Understanding factors influencing highway-rail crossing crashes is crucial for reducing accident rates and associated costs.
• The study provides valuable insights into railroad crossing safety applicable to various regions.