Using native vehicle observations to diagnose slick roads

The use of vehicle sensor data to improve weather and road condition products, envisioned as part of the USDOT Research and Innovative Technology Administration (RITA)'s IntelliDriveSM initiative, could revolutionize the provision of road weather information to transportation system decision-makers, including travelers. Vehicle-based probe data will likely increase the density of road-specific weather observations and also provide unique datasets for deriving and inferring road condition information.

In the spring of 2010, we conducted a field study in the Detroit, Michigan area to assess wireless vehicle observations from multiple vehicle types. Generally, the study targeted specific weather conditions in order to assess the accuracy of directly measured weather variables (e.g., temperature and barometric pressure). However, separate tests were performed in order to assess the utility of using other physical variables from the vehicles for suitability in pavement condition prediction. These variables included (but were not limited to) the horizontal acceleration measurements, yaw rate, and steering angle when the anti-lock braking system and the traction and stability control systems were engaged and NOT engaged. The results of this study showed a possibly significant change in the yaw, steering angle and longitudinal acceleration measurements when each of the aforementioned systems was engaged. However, the change in the latitudinal acceleration did not appear to be significant. These are very positive results because of the possibility that realtime indications of changes in these variables may provide useful information for future road slickness diagnostic products.

A more controlled study was also performed using four of the test vehicles and two friction measuring devices on a closed stretch of pavement under differing friction conditions. The results of this study showed similar agreement with the non-controlled study but were not conclusive. Further testing with multiple vehicles under an even more controlled environment will be required to garner a conclusive assessment of the usefulness of these particular vehicle observations and the possibility of diagnosing slick conditions on the roadways in an automated vehicle-based fashion.