Our paper presenting a synthesis of mathematical traffic models, observation models, and data assimilation techniques to incorporate Eulerian and Lagrangian traffic flow observations into a macroscopic traffic model is now available. The paper can be found online here.
- We develop and test methods to incorporate a variety of real-time traffic flow measurements from GPS, traffic sensors, traffic cameras, and cell phone data, etc. to predict macroscopic traffic flow.
- We show these methods lead to accurate predictions regardless of which data assimilation technique, ensemble Kalman or particle filter, is used.
- Both techniques allow for parameter estimation of difficult parameters to otherwise prescribe to the traffic models.