USDOT and Waze Partner to Predict Fatal Crashes
Taking a fresh look at its 750 "siloed and unrefined" national datasets, USDOT plans to identify and combat some of the less obvious contributors to fatal crashes—going beyond direct causes, like speeding, to proximate factors like time of day, road conditions, and weather.
As part of the new initiative, USDOT will partner with Waze to compare real-time and historical data on unsafe road conditions in order to gauge the likelihood of future collisions. Eventually, they hope to use artificial intelligence to find more nuanced connections among contributing factors.
Eno Transportation reports that the initiative will begin with two pilot programs:
The first project will help USDOT to better understand how vehicle speed, speed differentials, and roadway characteristics factor into fatal crashes. GPS-enabled devices will collect anonymized data on the prevailing speeds of vehicles across the entire National Highway System every 5 minutes. This will be the first time that USDOT has conducted such a study on a national scale.
The second pilot will leverage roadway hazard reports from Waze users on the road and USDOT’s historical crash data to estimate crash risks ahead of time. This might not only predict the likelihood of a crash, but also could present opportunities to mitigate those risks ahead of time by notifying drivers using the app.