Leveraging Connected Vehicles to Provide Enhanced Roadway Condition Information

Real-time performance measures are important for agencies to maintain their roadways during the winter season. However, infrastructure can be expensive to deploy and maintain and may be sparse in rural areas, while speed data alone may not provide enough fidelity in borderline conditions. This study looks at high-frequency vehicle data from the in-vehicle bus to detect changes in the vehicle and driver behavior during changing winter road conditions. The data is reported to the cloud via cellular communication and is viewable in real-time using a map-based web dashboard. Three winter weather events are assessed using in-vehicle data collected from the February-March 2018 period. Using a two-sample Kolmogorov-Smirnov test, we found high significance in the reductions of applied brake pressures and rates of braking in winter versus fair weather conditions before vehicle intervention is necessary. The paper concludes that pairwise comparison of driver change in brake pressure may be a valuable data source indicative of deteriorating road conditions before more severe indicators such as traction control, antilock brake, and/or hazard indicators are activated.