4B.3 The Truck Blowover Algorithm for the Pikalert® System

Tuesday, 9 January 2018: 9:15 AM
Room 10AB (ACC) (Austin, Texas)
Brittany Welch, Univ. of Illinois, Urbana, IL; and A. Anderson, S. Linden, and W. Petzke

Adverse weather conditions have an extreme impact on the ability to safely operate a motor vehicle. Over 5000 road weather fatalities occur yearly while road weather related delays can cost trucking companies over 2.2 billions dollars annually. These impacts are strongly felt along the 402-mile long Interstate 80 (I-80) corridor in Wyoming, where extreme weather occurs year-round at elevations between 6,000 and 8,600 feet. One of these extreme forms of weather, high winds, preferentially affects freight traffic due to their high profiles. Freight traffic makes up half of the daily traffic volume in this corridor; thus 50 percent of the vehicles traveling along I-80 are at high risk of being impacted by one of Wyoming’s most dangerous and common forms of road, wind. The Wyoming Department of Transportation (WYDOT), Federal Highway Administration (FHWA), and the National Center for Atmospheric Research (NCAR) are collaborating to combine vehicle data and weather data into applications that can aid in the drivers decision-making process when hazardous road conditions are present or expected as part of the WYDOT Connected Vehicle Pilot Deployment Initiative (WYDOT CV Pilot).

With previous support from FHWA, NCAR developed The Pikalert System, which is capable of ingesting, quality checking, and interpreting weather and vehicle data and outputting warnings, advisories, and roadway conditions. The Pikalert System currently uses three algorithms for hazardous weather advisories and warnings, but lacks a wind algorithm, which is vital along the I-80 corridor.

To address this shortcoming, a blowover algorithm was developed and added to the Pikalert System using a fuzzy logic methodology. A small group of case studies were used to assess the initial results of the algorithm, and based off of these results the weights and functions within the algorithm were changed to provide a more appropriate interest value output for each case. The algorithm was further verified against a seven year I-80 crash data set. The resulting algorithm, along with the rest of the Pikalert System, will be available to the community as part of the open source code developed for the WYDOT CV Pilot.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner