The increasing availability and implementation of in-vehicle data recorder (IVDR) technologies has resulted in an increased number of safety applications based on the behavior of drivers during actual driving. The objective of this study is to formulate a flexible framework for IVDR data analysis that has the capability to include roadway geometric attributes, driver characteristics, environmental factors and time-dependencies. This research uses the concept of a homogeneous trip segment (HTS), applied with a survival model to estimate the risk of an event occurrence (e.g. crash or warning from on-board monitoring system). The University of Michigan Transportation Research Institute (UMTRI) provided data from the Road Departure Crash Warning System Field Operational Test. The survival-based model allowed the assessment of event hazard for individual drivers on specific routes as well as individual drivers compared to their peers, while controlling for travel in different routes and under different conditions. The authors believe this capability offers opportunities to safety researchers that were not available from existing methods.
|Journal||Transportation Research Part C: Emerging Technologies|
|State||Published - Apr 2021|
- Behavioral-based safety
- Crash risk modeling
- In-vehicle data recorder (IVDR)
- Time-varying coefficients
- Time-varying effects