Laoufi, M., Blum, J., and Eskandarian, A. (2004) “Predicting Intersection Queues with a Dynamic Balance Model”, San Antonio, TX: Intelligent Transportation Society of America 2004 Annual Meeting, CD-ROM.

Abstract: The ability to accurately predict intersection queue length is a crucial input for adaptive signal timing systems and route planning applications. One way to obtain accurate measures is through the costly deployment of additional sensors. This paper introduces a Dynamic Balance Model, which makes predictions without requiring additional sensors. The Dynamic Balance Model specifically overcomes limitations in the data provided by typical sensor configurations through the application of conditional probability measures. The model was tested in realistic simulations of arterial roadways in Northern Virginia, and produced extremely accurate results.