The challenge of obtaining ground truth for GPS processing

Abstract

The increasing use of GPS as a substitute or complement to conventional travel surveys has brought with it an increasing need for a reliable source of ground truth, i.e., information on the actual travel in which each respondent engaged, including the mode and purpose of such travel. Over the past decade or so, the main source that has been used for providing ground truth is the prompted recall survey. As the survey task has become less of a challenge, it has become evident that the prompted recall survey is subject to many of the same shortcomings as most selfadministered surveys, with reporting error, misunderstanding of what constitutes a trip, etc. In this paper, we review some of the common problems encountered in prompted recall surveys, especially those undertaken at some period of time after the GPS survey was undertaken, and using self-report methods. Following this, the paper describes a recent experiment in using life-logging cameras to record a person’s travel, together with GPS, thereby providing a new source of ground truth data that is not subject to processing issues of the GPS data, nor to self-report or other issues with prompted recall data.