Network-Wide Traffic Volume Estimation Based on Probe Vehicle Data

The authors propose a method to estimate traffic volumes on all roads of the network, relevant to traffic use cases. Using large scale probe vehicle data in combination with stationary detector data, the method builds upon a model to estimate the probe vehicle penetration rate on road level. Exploratory data analysis has shown that the average penetration rate measured in the vicinity on the same road class, the province identifier and the average historic probe vehicle volume have the strongest predictive power on the penetration rates of a road. Trained by penetration rates from only 60% of the stationary detectors in the Netherlands, the model is able to predict the probe vehicle penetration rate with a Mean Absolute Percentage Error (MAPE) of 8.3%, and decreases to 6% on motorways. Increasing the proportion of detectors used for training did not significantly improve model performance. This suggests the method can be used in countries with less developed detector infrastructure as well. Using the estimated penetration rates, traffic volumes on motorways can be estimated with a MAPE of 25% already within about two minutes of collecting probe vehicle observations, and of 7% within a day. Thus, the traffic volume estimations can support real-time operations on motorways. For roads other than motorways, the maximum accuracy is limited to a MAPE of about 12%, and significantly longer observation periods are needed. Still, most offline use cases can be covered, e.g. annual average daily traffic.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: References;

Subject/Index Terms

Filing Info

  • Accession Number: 01908539
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-24-00825
  • Files: TRIS, TRB
  • Created Date: Feb 20 2024 9:16AM