Conference Interactive Program
To view the 2013 Annual Meeting Interactive Program
TRB 90th Annual Meeting (January 23-27, 2011)
Intelligent Transportation Systems Development and Applications
Jan 24 2011 7:30PM- 9:30PM
Marriott, Salon 2
Real-Time Freeway Experienced Travel-Time Prediction Using N-Curve and k Nearest Neighbor Methods (11-4060)
Essential for the development of efficient traffic control and management strategies is the dissemination of reliable traffic information. Broadcasting travel information on Intelligent Transportation Systems (ITS) components, e.g., dynamic message signs (DMS), is of benefit as travelers are notified of near future traffic conditions allowing them to make informed travel choices based on traffic behavior along the network, overall improving network performance. Such results prove ITS and its disseminated information as indispensable. This study presents simple and reliable models for short-term freeway travel time prediction. The proposed models are founded on the developed input-output (N-Curve) method. In these models, the N-Curve method is extended to account for ramps, a more realistic situation, when compared to the originally proposed. The prediction model utilizes the N-Curve method in conjunction with the k-Nearest Neighbor (k-NN) method to predict experienced (or actual) travel times. A real-world based traffic network and demand are replicated in a traffic simulation model where several scenarios are produced to serve as the test bed for algorithm evaluation and validation. The proposed Single-NN algorithm best predicts travel times for light, free flow traffic conditions while the Multiple-NN algorithm best predicts travel times for peak period traffic conditions. The Hybrid-NN algorithm successfully merges the Single-NN and Multiple-NN algorithms, exploiting each algorithm where most suitable. Light traffic conditions exhibit RMSEs results for the Hybrid-NN algorithm below 6% and 45% for the westbound and eastbound direction, respectively. Moderate traffic conditions for the RMSEs are below 6% (westbound) and 31% (eastbound) while RMSEs for heavy traffic conditions are RMSEs are below 6% (westbound) and 28% (eastbound).
Bustillos, Brenda , University of Arizona
Chiu, Yi-Chang , University of Arizona
Transportation Research Board. 500 Fifth St. NW, Washington, D.C. 20001
Copyright © 2012. National Academy of Sciences. All Rights Reserved.