Citation

Jiang Y. J. Transp. Res. Forum 2003; 57(2): 99-118.

Copyright

(Copyright © 2003, Transportation Research Forum)

DOI

unavailable

PMID

unavailable

Abstract

This paper presents a method for predicting freeway traffic flows that combines the Klaman control theory and the time series theory. It is illustrated that the combination method provides more accurate traffic flow prediction than using either one of the two theories individually. With the prediction model, the traffic flow on a given freeway in the next time interval (5 to 15 minutes) can be predicted using traffic data at the current and past time intervals. Dynamic traffic predictions with the developed model can be performed for individual lanes as well as for all the lanes of each travel direction. The study also demonstrates that, if the traffic capacity was a given, a dynamic prediction of traffic flow rate with this prediction model would also constitute a dynamic prediction of traffic congestion.