1. Home
  2. TRID
  3. View Record
https://nap.nationalacademies.org/catalog/27432/critical-issues-in-transportation-for-2024-and-beyond

Mode detection algorithms for GSP data

Kulkutavanpaattelyalgoritmit GPS-aineistosta

The purpose of this study was to find out what kinds of possibilities GPS has to offer for travel surveys. The study focused mainly on household travel surveys and, more specifically, the detection of mode of transport from GPS data. The results can also be applied to other kinds of traffic and travel surveys where the reasons behind travel, methods of transport and characteristics of trips are studied. Traditional methods, such as interviews and travel diaries, have their limitations. Answering all the questions is often a time-consuming task for respondents, response rates are falling, some trips are forgotten altogether, and estimates about travel distances and durations are inaccurate. Because of these weaknesses, new methods have been developed. Their goal is to automatize the data collection and thus facilitate participation in studies, enhance data quality and, if possible, also to reduce costs. Studies that use new methods can also involve longer periods of time and more people. Using GPS devices is but one of these new methods. First, relevant literature was examined to find out what kinds of possibilities GPS offers to travel surveys, what limitations there are to the technique and what kinds of surveys have already been carried out. In addition, automatized processes for detecting information needed in travel surveys - mode, purpose of trip, route - from GPS data were examined. The focus was especially on algorithms for detecting the mode of transport from GPS data. The study included a pilot study of 50 persons carrying a GPS device to gather travel data. They also kept an extensive travel diary including times and modes of transport used. Based on this data, a mode detection algorithm was developed. The quality of a mode detection algorithm depends on what modes need to be distinguished and the information available on the traffic system. The best algorithms can detect the mode correctly in 95 percent of trips. In this study, the result was around 80 percent. The modes detected were walking, bike, car, bus, tram, metro and train, and geographic information about public transport stops was used in addition to the GPS data. It is possible to improve the results, for example by using more geographic information such as public transport routes. Using GPS is a promising method, but it cannot be used to gather all the information that is nowadays gathered in household travel surveys. Thus GPS cannot fully replace the methods currently in use, but can only complement them. Using GPS also brings new risks to the surveys. The costs and the quality of data have to be assessed beforehand. This report may be found at http://alliohjelma.info/dokumentit/julkaisut_sarja/

Language

  • Finnish

Media Info

  • Pagination: 157p
  • Serial:
    • ALLI-JULKAISUJA
    • Issue Number: 6/2010
    • Publisher: MINISTRY OF TRANSPORT AND COMMUNICATIONS

Subject/Index Terms

Filing Info

  • Accession Number: 01321575
  • Record Type: Publication
  • Source Agency: TRL
  • ISBN: 978-952-255-015-6
  • Files: ITRD
  • Created Date: Nov 29 2010 11:40AM