United States Department of Transportation - Federal Highway Administration FHWA HomeFeedback
OHPI Logo Office of Highway Policy Information

Measuring Day-to-Day Variability in
Travel Behavior Using GPS Data



This chapter provides a brief overview of the GPS-based travel survey experiment that was conducted by Federal Highway Administration (FHWA) in Lexington, KY. Much of the material in this chapter is derived from a report prepared by Battelle [1997] that includes detailed descriptions of the experiment and the databases assembled.

Selection of MPO

The GPS-based travel survey experiment required the participation of an MPO. The selected MPO would need to have an up-to-date and accurate digital map of the test area and its highway network, with a minimum accuracy satisfying Federal National Map Accuracy Standards. In addition, the participating MPO would have to be able to provide about 250-300 hours of total staff support during the field experiment.

Following a nationwide solicitation and selection process, the Lexington area MPO was chosen as the host MPO for the field test. The Lexington area MPO is the principal planning agency for the two-county area of Fayette and Jessamine in Central Kentucky. The two counties together encompass an area of approximately 461 square miles with a total population of approximately 350,000.

Recruitment of Households

Considering the specialized nature of the GPS-based field experiment, it was not possible to utilize a pure random sample for the survey. A special recruitment process was used to obtain a final sample of 100 participating households for the experiment.

A sample of nearly 2,000 listed telephone numbers from households in Fayette and Jessamine counties (weighted by population) was purchased from a commercial source. The participating households were recruited using a stratified sampling plan based on several demographic factors. The demographic factors considered included gender, age, and presence of children in the household. Licensed drivers under the age of 18 were not permitted to participate as the principal eligible driver within a household. In addition, principal eligible drivers recruited for the field test were required to drive at least three days a week. This was done to ensure that a reasonably large amount of data would be collected despite the small sample size and short duration of the experiment.

The recruiting process started by mailing out pre-solicitation letters without any prior contact with the households. Once the letters were mailed, telephone contacts began with the households. If the household was responsive to the initial call, they were asked to participate in a brief screening survey to determine their eligibility for the field test. Following the screening interview, a pre-usage interview was conducted to ensure that the household would be able to use the field equipment without any problems.

A pre-solicitation letter was sent to a total of about 1,300 households with listed telephone numbers. Once the telephone interviewers determined that there was an eligible driver and vehicle in the household, 67 percent of those eligible consented to participate in the field test. At the end of the recruitment process, a total of 100 households were successfully recruited for participation in the field test.

Field Equipment

The field equipment consisted of the following individual components:

Personal Travel Survey Software

The Personal Travel Survey (PTS) software had two primary functions. First, it allowed respondents to easily enter trip information and second, it facilitated recording of GPS positional data for each trip. It consisted primarily of three interfaces - administrative, GPS, and respondent.

The administrative interface allowed the field test administrator to set the operational parameters of the data collection devices and personalize the respondent interface for each household. Household users were not provided access to this interface.

The GPS interface displays raw GPS output as it is being written to memory in the computer. This interface provides the mechanism by which the GPS receiver data stream is written and stored in memory. This interface does not have any user control.

The respondent interface is the major feature of the PTS software that respondents control and utilize throughout the field experiment. The interface allows drivers to use a touch-screen to initiate a trip, identify the driver and his or her activity (trip purpose), identify passengers and their activities (trip purposes), and end a trip. Note that time, speed, and positional data are recorded directly through the GPS receiver.

In general, both the hardware devices and software interfaces performed well in the field. There were very few instances of hardware or software malfunction that called for corrective action. Further details about the hardware and software performance can be found in Battelle [1997].

Back to Main Page

FHWA Home | OHPI Home | Feedback

United States Department of Transportation - Federal Highway Administration