ITS Research Program, UW

Cellular Automata:
A Cellular Automata Model for Use with Real Freeway Data for Congestion Prediction

Sponsor:

Washington State Department of Transportation

Description:

This project will extend and validate a cellular automata model using freeway traffic data form the I-5 corridor in North Seattle. When the model is accurate enough that it reproduces the observed statistics of traffic behavior, the model will be tested in predictive mode. The predictions from this model will be quantitatively compared to the observed freeway time series both temporally (predictions of the future) and spatially (predictions of down-stream statistics). The scale of the prediction problem is 15 miles downstream and approximately 15-20 minutes ahead.

Cellular automata are mathematical idealizations of physical systems in which space and time are discrete, and physical quantities take on a finite set of discrete values. They were originally introduced in the 1960s with the particular purpose of modeling biological self-reproduction. Since then, the theory has been used broadly in physics for particle transport simulations and thermodynamics studies.

The Seattle ITS Backbone makes large amounts of real-time data available, and the TDAD data mine contains almost one year of 20-second data at the single loop/lane level for the Puget Sound metropolitan region. These two data sources, combined with past work, make this a nationally unique opportunity to develop this technology. This work will open new arenas of prediction for ATIS applications and new avenues for planning activities. It will be possible to predict both the effects and clearing time of incidents and, therefore, to determine the appropriate distribution of resources for incident response teams.

References:

"Real-Time Highway Traffic Simulation and Prediction Using Inductance Loop Data," X. Hu and D.J. Dailey, Proceedings of the Vehicular Navigation and Information Systems 1999, pp. 194-9, Seattle, Washington, 20 July-2 August 1995.