Survey Information
Methodology
Aerial photography was selected as the best means of collecting mobility data across the large Atlanta planning region. First, aircraft move rapidly and are not slowed by congestion on the highway, making for a low per-mile cost of data collection. Second, the aerial point-of-view provides for photography that clearly shows not only how traffic on the highways is flowing in both directions, but also reveals the underlying cause of most congested bottlenecks; it also shows the presence or absence of congestion on crossing streets. Third, use of a remote sensing methodology means that no equipment needs to be installed on the ground that interferes with the movement of traffic, or exposes workers to the dangerous highway environment. This also means that data can be collected along any route, regardless of whether the route has been equipped and wired with ground-based traffic sensors.
During survey operations, 100% overlapping aerial photographs are acquired of all surveyed highway links at a frequency of one sample per hour, repeated weekdays between 6:30 and 9:30 a.m., and between 4:00 and 7:00 p.m. So that the effects of incidents can be identified and removed, and to reduce the effects of daily fluctuations, coverage is repeated over four different weekday mornings and evenings. After flights have been completed, photography is segregated by highway, and counts are made of the number of vehicles operating in each segment, in each direction. Raw data are entered into a database file for computation of densities, levels-of-service, and surrogate levels-of-service (the latter for interrupted-flow highways).
Once performance measures have been computed for all flights, data scrubs are conducted to screen for anomalies by comparing the results, segment by segment, across different days and time slices. This process is used to identify the effects of incidents and other non-recurring events that affect traffic quality. Where confirmed or suspected incidents / events are found, corresponding data are tagged for exclusion from the averaging process. Final (averaged) performance ratings, therefore, reflect normal traffic conditions, without the effects of non-recurring events. It is these ratings that form the technical basis for the analyses in the survey deliverables.