Project Number:
04-15
Research Project:
Confidence Intervals for Estimated Traffic Demand
P.I. Name & Address:
Fernando Ordonez
Department of Industrial and Systems Engineering
University of Southern California
Los Angeles, CA 90089-0193
Tel:(213) 821-2413
Fax:(213) 740-1120
Email: fordon@usc.edu
Co-P.I.s:
Kurt Palmer
Department of Industrial and Systems Engineering
University of Southern California
Los Angeles, CA90089-0193
Tel:(213) 740-5960
Fax:(213) 740-1120
Email: kpalmer@usc.edu
Project Objective:Objective:
We propose to develop methodologies to obtain confidence intervals on estimates of the traffic demand in a transportation system.The objective is to construct confidence intervals on the demand, represented through origin-destination (OD) flows that depend only on the uncertainty present in the data used to estimate this demand. We also propose to conduct an estimate of OD flows for the Los Angeles highway system, using data that is routinely collected from the highway system and publicly available demographic information.
Abstract:
The transportation demand, in the form of origin-destination flows, is a key input to diverse models that address many of today's transportation problems.For example models that use OD flows are present in reports issued for planning purposes by institutions such as the MTA and SCAG; and OD flows are central on Dynamic Traffic Assignment models that are used for the analysis and management of urban and suburban congestion problems.The effectiveness of these models is influenced by the accuracy with which the traffic demand can be represented.Furthermore, there is an inherent uncertainty associated to the traffic demand, as population's driving habits change, any measuring/estimating method includes some level of error, and the extrapolation of the sampled data to the planning horizon causes additional inaccuracies.
The use of confidence intervals on the estimated OD flows addresses two difficulties faced by current approaches to model traffic demand: 1) it provides a model to represent the uncertainty present in the demand and 2) it does not involve making arbitrary distribution assumptions on traffic demand.Additionally, the confidence level of a confidence interval can be used by the planner as the means to compare risky versus risk-averse solutions to the problem through the same model.A better understanding of OD flows improves our understanding of the transportation system and, in time, can lead to better solution methods and algorithms.
This proposal considers the study, development and creation of confidence intervals for traffic demand in the form of origin-destination flows.The objective is to obtain methods that provide confidence intervals that depend on the uncertainty present in the data used to estimate the OD flows, and can thus be measured and calibrated for different applications.We expect to produce confidence intervals for origin-destination pairs for a moderate sized model of the Los Angeles highway system.
We will build on existing models for OD flows estimation from observed traffic flows.Our research will investigate the conditions under which the OD estimates, obtained through Kalman filters, provide enough information to construct confidence intervals.These confidence intervals will be contrasted with those obtained through the direct repetition of an estimation model
Task Descriptions:
1. Literature review of current OD estimation methods (3 months)
2. Development of methodology to construct confidence intervals (5 months)
3 Data gathering and model definition (5 months)
4. Construction of OD confidence intervals (2 months)
5. Preparation of final report (2 months)
Milestones, Dates:
January 5, 2004 -January 4, 2005; final draft report 5/31/04
Total Budget:
$75,000
Student Involvement:
One Student @ 50% effort 12 months
Relationship to Other Research Projects:
Related to 99-25, part of urban mobility focus area
Technology Transfer Activities:
Project report posted on the website
Potential Benefits of the Project:
Better transportation demand modeling
TRB Keywords:
traffic demand; networks
Primary Subject:
4b.1 Measurement characterization and modeling of system performance and impact measurement
Goals:
4c.2 Mobility
Enabling Research:
4c.11 Tools for modeling and design
Modal Orientation:
Highway
