Skip Top Navigation

Research Reports

Automata Model for Congestion Prediction

Description: The overall goal of this project is to create a method for predicting traffic congestion on freeway corridors. When implemented, it will provide a traffic service like that of “pin-point Doppler” weather radar that can predict growing or dissipating congestion. Preliminary versions of the model used real-time loop data to successfully reproduce traffic behavior under moderately congested conditions. To improve the model for heavily congested conditions, the model had to be accurately calibrated. The process of calibrating the model revealed that inductance loop errors were preventing accurate results. An algorithm to correct the data from improperly functioning loops was developed and published at the 2003 Transportation Research Board Annual Meeting. The corrected loop data from the TDAD (Traffic Data Acquisition and Distribution) data mine are now being used to calibrate the model. The algorithm created in support of this effort can be used with malfunctioning loops to improve the freeway management system performance monitoring effort.

  • Date Published: December, 2003
  • Publication Number: WA-RD 577.1
  • Last Modified: April 23, 2007
  • Authors: Zachary Wall, Daniel J. Dailey.
  • Originator: Washington State Transportation Center (TRAC)
  • # of Pages: 73 p., 4,339 KB (PDF)
  • Subject: Algorithms, Calibration, Errors, Freeways, Loop detectors, Mathematical models, Mathematical prediction, Real time data processing, Traffic congestion, Traffic data.
  • Keywords: General automata, loop data, real-time data, traffic management, congestion prediction, recurring congestion, non-recurring congestion, car following.
  • Related Publications:


This abstract was last modified April 29, 2008