Development and implementation of a fuzzy logic-based blowover algorithm

Wyoming experiences high winds that pose a persistent threat along the Interstate 80 (I-80) corridor. These conditions increase risk to freight traffic, making truck blowovers the predominant type of non-collision crash in Wyoming and result in roadway closures, economic losses, and fatalities. Blowover prevention using decision support systems was a focus during the Wyoming Department of Transportation’s participation in the Connected Vehicle Pilot Deployment. One decision support system used was the Pikalert® system, developed by the National Center for Atmospheric Research. Pikalert combines vehicle and weather data to generate weather-hazard assessments but lacked an algorithm for blowovers. This paper reviews the development and implementation of such a blowover algorithm. A literature review identified important weather and road geometry factors in blowover crashes. The developed algorithm incorporated wind gust speed, differences between sustained and gust wind speeds, wind direction relative to roadway orientation, and roadway slickness. Functions, thresholds, and weights were informed by prior blowover research and empirically refined through crash case analyses, sensitivity testing, and real-time verification during the 2017–2018 winter season. Wind gust speed and differential were the most influential parameters across vehicle classes, as sudden lateral forces destabilize high-profile vehicles. The truck blowover algorithm remains an active part of Pikalert, which is available to multiple transportation agencies. The broader impact and merit of this work is to support transportation agencies for preemptive roadway closures and vehicle restrictions during high wind conditions, thus reducing blowovers, roadway closures, and mitigating road weather related economic losses.