|
Activity Number:
|
250
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #310380 |
|
Title:
|
Estimating Traffic Volume over a Large Transportation Network from Noisy Data
|
|
Author(s):
|
Jaimyoung Kwon*+ and Karl Petty
|
|
Companies:
|
California State University, East Bay and Berkeley Transportation Systems, Inc.
|
|
Address:
|
Dept of Statistics, Hayward, CA, 94542,
|
|
Keywords:
|
network ; graph theory ; flow conservation ; weighted least squares ; space-time data
|
|
Abstract:
|
Accurate measurement of traffic volumes throughout a transportation network is critical for efficient traffic management and as well as for simulation studies. Obtaining a network-wide volume estimate is a challenging task since: a) transportation networks are large and complex; b) measurements from many sensors, each with different noise level, need to be fused together; and c) the final estimate needs to globally satisfy the flow conservation constraint. We develop a graph-theoretic, weighted least squares regression (WLSR) approach for the problem. The approach uses graph theory to construct the design matrix for WLSR, which represents the flow conservation equation. Also, measurement errors are reflected a priori as the weight matrix for WLSR. The performance of the algorithm is studied via simulation and application to data from real freeway networks in Greece and California.
|
