Traffic Prediction

Learn More

Traffic information supply about the traffic state in urban and metropolitan areas is a priority both for public entities and business for which such an information is a key factor. The relevance of the traffic web page in most city council sites or the rising quantity of mobile apps and software applications are a clear indicator. Never the less a minimum analysis of any of them immediately shows that:

  • Reliable information is limited to that from main arterials, where sensors have been placed, and that the rest of the streets have no reliable information or no information at all
  • The available information is worthless for navigational systems, route calculation based on travel times, etc
  • The backend has no short term predictive capabilities, that is that the information provided is directly obsolete (for instance information about the travel time of vehicles that have already completed their trip rather than information about how long will it take from the present moment)

As a consequence the real usefulness of those traffic information systems is very limited, in most occasions restricted to something as quick visit before starting a trip and, less frequently, along the trip as a support for the navigational system.

Betterways system for the estimation of the traffic state and the prediction of its short term evolution has been designed and developed to cover these deficiencies. It counts with two components:

  • A module for the estimation and short term prediction of mobility patterns (Origin-Destination matrices, OD) that exploits the versatility and soundness of the models based on Kalman’s filter combined with the mix of traffic measurements provided by ICT (especially the supply of travel times)
  • A module for Data Completion based on the mesoscopic simulation of the temporal evolution of traffic flows, which generates complete information about the traffic state and its short term evolution throughout the whole network. Information generated through simulation is consistent with the one generated by the traffic sensors