Development of Congestion Prediction Models - Phase 3

Under past research, a model is under development at the University of Washington (UW) and is being tested with real roadway data from I5 and SR168. The scale is 15 miles downstream or 15 to 20 minutes ahead. The model can predict recurring congestion but needs to be expanded to model non-recurring incident created congestion and the dissipation of that congestion. Completion of a model that accurately predicts future congestion based on the physics of traffic will provide a foundation for real-time optimal control and a test bed for modeling roadway capacity changes such as lane closures and incidents.