Paper Details

TRB 91st Annual Meeting (January 22-26, 2012)

Event Number:	594
Event Title:	Harnessing New Data Sources: Bluetooth Readers and Probes
Event Date:	Jan 24 2012 2:00PM- 3:45PM
Event Location:	Hilton, International Center
Event Description:
Event Agenda:

Can Historical Speed Data from Private-Sector Sources Be Used to Flag Segments with Congestion on Interrupted-Flow Arterial Highways? (12-4710)

Transportation agency operators and planners can now purchase real-time or historical vehicle speed data from private-sector vendors covering very large regional highway networks. The accuracy of these data products is fundamentally important; however, operators and planners view “accuracy” through different lenses, and while freeway data quality has been comprehensively studied, authoritative conclusions about suitability of these sources for interrupted-flow arterial monitoring have not yet been published. Thus, working from a planner’s perspective, the author has taken a macro-level first-cut look at the question. For test highway segments where he already knew where typical daily bottlenecks were located, he calculated typical peak-period delay values – segment-by-segment – from default “historical” speed data donated by a private vendor. He juxtaposed these delay calculations against his lists of confirmed daily bottlenecks – the sources for these lists were periodic historical surveys using large sets of time-lapse aerial photography (TLAP), acquired during 24 mid-week morning and evening flyovers. He then refined calculation rules and experimented with various threshold delay values (above which segments would be flagged as “congested”), until he found the best correlation between “congested” ratings generated by the two methods. Reviewing these lists, he found evidence that private-sector speed data on interrupted-flow arterials may already be good enough to classify segments as “congested”, with varying degrees of severity and duration. After more rigorous research to confirm, planners may have a more affordable way to calibrate computer models, screen large networks for likely bottlenecks, confirm where long-term degradation is occurring, or prove the benefits of investments.

Authors
Jordan, Gregory Wayne, Skycomp, Inc.