Road Work Ahead: Using Deep Neural Networks to Estimate the Impacts of Work Zones

Roadside construction – be it a detour, a closed lane, or a slow weave past workers and equipment – work zones impact traffic flow and travel times on a system-wide level. The ability to predict exactly what those impacts will be, and plan for them, would be a major help to both transportation agencies and road users. Funded by the National Institute for Transportation and Communities, the latest Small Starts project led by Abbas Rashidi of the University of Utah introduces a robust, deep neural network model for analyzing the automobile traffic impacts of construction zones.