Advanced Driver Assistance System Crash Rate Assessment Using Vehicle-Specific Mileage Data

Leveraging vehicle miles traveled (VMT) data provides a feasible means for researchers to control for exposure when evaluating the effectiveness of advanced driver assistance systems (ADAS) in reducing system-relevant crashes. The latest study by the Partnership for Analytics Research in Traffic Safety (PARTS) expands on prior research that pairs auto manufacturer equipment with police crash report data covering 98 million vehicles and 21.2 million crashes.

For the first time, this PARTS study included VIN-specific VMT data for two vehicle segments (small sport utility vehicles and pickup trucks) collected in one of two ways: during dealer visits and via vehicle telematics. Nearly 300,000 vehicles in the dataset have VMT data, equipage information, and crash data, which enabled this enhanced study of ADAS effectiveness.

The report shows that, despite variability in manufacturer reporting practices and changes over time, VMT data produced results consistent with previous studies, highlighting its reliability for traffic safety research.

The study also reveals that more frequent VMT readings improve accuracy, but estimates remain reliable with time gaps of several months. Aggregated measures, such as annual or monthly mileage per vehicle, are sufficient for certain types of safety analyses. This approach avoids the complexity and volume of sensitive trip-level data while maintaining analytical effectiveness. The report underscores the importance of establishing standardized reporting protocols to enhance the utility of pooled datasets.

Integrating telematics data represents a pivotal improvement to PARTS’ ability to assess ADAS effectiveness. Future enhancements will expand the use of telematics data to improve traffic safety research capabilities and better understand the state of the vehicle before, during, and after safety critical events.