Use of Multisource Global Positioning System Data to Characterize Multiday Driving Patterns and Fuel Usage in a Large Urban Region

  • Venter C
  • Joubert J
  • 16


    Mendeley users who have this article in their library.
  • 2


    Citations of this article.


in a conventional travel diary survey and to calculate correction factors (2)—to a full-sample survey instrument. A number of large-sample household travel surveys that use GPS (with or with-out prompted recall surveys) have been concluded or are under-way in, for instance, Cincinnati (1,500 households) and Jerusalem (3,000 house holds) (3, 4). At the same time, data collection periods are increasing beyond the traditional 1-day survey; respondents have carried GPS devices for up to a week and provided rich data on the day-to-day repetitiveness and variability of travel without incurring undue respondent burden (5, 6). Compared to standard travel surveys, GPS-recorded personal movement data are seen as beneficial in terms of enhanced accuracy (both of locations and routes), completeness, and efficiency (2, 5). In-vehicle GPS devices and automatic vehicle location applica-tions have long been used to monitor and manage commercial vehi-cle fleets (mainly buses and trucks); these devices provide another potential source of GPS movement data (7). Despite the availability and size of such data sets, they have only recently been explored as a source of truck movement data for public agencies (8). Commercial GPS data could significantly advance a disaggregate understanding of the behavior of road freight traffic (9). As far as the authors are aware, GPS data from commercial fleet management and project-specific mobile GPS sources have not yet been combined to analyze and compare private and commercial vehicle traffic in the same area. The potential efficiency of GPS technologies, coupled with their ubiquity and flexibility, make them attractive for the collection of travel data in developing countries. GPS surveys might help over-come some of the problems with literacy and respondent burden associated with conventional activity or travel diary surveys in less literate or multicultural populations (10, 11). Relatively few studies, however, have assessed these issues outside of the developed world. A recent review of GPS travel studies worldwide found no applica-tions (in the English literature) in South America or developing Asia (12). Some experimentation has been done in the Western Cape, South Africa, to compare the results from wearable GPS data with those from conventional travel diary data and to demonstrate the usefulness of the GPS data in areas with limited address and street network information (13). This paper reports on the use of GPS data to estimate fuel con-sumption at a very disaggregate level (by link and time of day) for multiple types of vehicles in the Gauteng urban region in South Africa. Fuel consumption is linked with the observed travel behav-ior of private (car) users and commercial (freight) vehicles and provides insights into the travel behavioral factors that affect fuel consumption. The paper describes the use of Global Positioning System (GPS) data obtained from commercial and project-specific sources to examine the travel behavior and fuel consumption patterns of drivers over a 3-day period in Gauteng Province, South Africa. Data for commercial (truck and light delivery vehicle) traffic were obtained from a commercial fleet management provider that continuously tracked the movements of 42,000 vehicles. Data for private car users came from a panel of 720 drivers, whose multiday driving activity was tracked by mobile pas-sive GPS loggers. The driving behavior of the two driver populations was analyzed and compared in terms of the total distance traveled: spa-tial patterns (e.g., the amount of travel on different types of road) and temporal variations (e.g., variations between times of day and between multiple days). The detailed nature of the GPS data permitted the esti-mation of fuel consumption at a very disaggregate level (by link and time of day) and the identification of differences between user groups; these factors have significant implications for transport and energy policy. A new indicator, the recovery ratio, was introduced to assess the relationship between fuel use and distance traveled on different classes of road and to help identify equity distortions between user groups. The paper also discusses research needs related to the collection and inte-gration of GPS data from multiple sources for model calibration and program evaluation. The use of technologies based on Global Positioning Systems (GPS) to collect travel data is growing fast worldwide. As GPS tech-nologies improve, and as research and practical experience grows, a number of shifts are occurring in the application of the technologies. Technological improvements in the cost, weight, and battery life of mobile GPS logging devices are causing a shift from vehicle-based to person-based measurements of mobility. Wearable GPS studies have been conducted worldwide, including in the United States, Europe, Japan, and Australia (1, 2). Wearability, in turn, is shifting the use of GPS data from a supplementary source of travel data— typically used to audit the trip rates of a subset of the respondents 2 Transportation Research Record 2338 The work is significant for its international focus: specific issues related to the collection and use of GPS data in developing coun-tries are reported. Second, the feasibility of combining GPS data from two independent sources—a small-sample study that used mobile GPS devices and commercial fleet management GPS data— is demonstrated to provide a more complete picture of travel-related resource consumption in the area. It is argued that the high quality data obtainable from GPS sources—including information on multi-day travel patterns, detailed route choice and speed, and day-to-day variability—can improve fuel use studies beyond the information obtainable from traditional survey or modeling sources. Some of the problems and requirements of combining GPS data from mul-tiple sources are also discussed. The paper's third contribution is substantive: the analysis of fuel use patterns delivers insights into the differences in driving behavior between different user groups and in those groups' respective contributions to the overall fuel bill. The policy implications of these findings are discussed in relation to fuel efficiency programs and road funding, and recommendations for further research are made. The paper starts by discussing the public policy and technical antecedents for this work. The objectives and methodology for the collection and analysis of the respective sets of GPS data are then described, and data collection issues that might be relevant to other developing countries are considered. Third, selected findings about the travel behavior of private and commercial vehicle drivers in Gauteng are presented, followed by a description of the fuel con-sumption model and its results. The paper concludes with findings and recommendations for further research. Background

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document


  • Christoffel Venter

  • Johan Joubert

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free