[SPACKLE]
An interdisciplinary research project from the Data and Information Management Laboratory and the Intelligent Transportation Systems Laboratory.

People

PI:David Maier
Co-PI:Kristin Tufter
Robert Bertini
Rafael J. Fernández-Moctezuma

Project description

Many products and services in an Advanced Traveler Information System (ATIS) require complete and current monitoring data in order to provide timely and useful information to travellers. An example of such an ATIS product is TripCheck, a tool provided by the Oregon Department of Transportation for accessing traffic information through the Internet before (or conceivably during) a commute. In particular, it presents a color classification of current average flow speeds for various segments of the highway system. TripCheck empowers commuters to make informed decision concerning commute plans.

Inductive loop detectors deployed at various mileposts along highways and fr. While loop detectors remain a popular choice for gathering traffic data, they are known to fail for a variety of reasons. Failure in sensors may appear as total loss of data or noisy data. Incomplete or unreliable measurements have been noted as undesirable for data archiving systems and final products relying on archived data, leading to various techniques to impute values to replace missing or defective values in archived. We also note that missing values cause problems for products based on live data. The figures below shows a snapshot of TripCheck with missing information in two segments (circled.)

[TripCheck speedmap 1 [TripCheck speedmap 2

With such an information product, we conjecture that providing an estimated system state may be better than displaying incomplete or erroneous data. One approach to obtain more complete information on a segment state consists in building estimators that look at nearby sensor information for regression. Preliminary results are promising and encouraging, motivating further studies on this approach to address the issue of incomplete data presentation in ATIS information products.

This research will make use of The Portland Transportation Archive Listing (PORTAL), the Portland metropolitan area's transportation data archive. PORTAL has been archiving speed, volume, and occupancy data from the loop detectors on Portland area freeways since July 2004. PORTAL receives a live stream of freeway loop detector data from the Oregon Department of Transportation. This stream consists of 20-second volume, occupancy, and speed measurements for nearly 500 freeway detectors and 128 metered on-ramps in the metropolitan area. In addition to loop detector data, PORTAL archives weather data and freeway incident data. PORTAL provides a web-based interface, giving easy access to both raw data sets and a wide range of common summary data and performance measures. Currently, PORTAL downloads detector data on a daily basis. We expect during the course of this project that PORTAL, or an ancillary service, will also provide near-real-time access to detector data, via the TransPort Advanced Traveler Information Implementation (TATII). TATII is an ATIS infrastructure that allows Portland regional transportation agencies to share data in a standardized format to each other and to traveler information service providers over a high-speed network.

Products

Sponsors


Rafael J. Fernández-Moctezuma
Last modified: CDT

[ CS | CECS | PSU | Home ]