Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
DATA- ORIENTED TRAVEL BEHAVIOR ANALYSIS BASED ON PROBE PERSON SYSTEMS Eiji Hato and Ryuichi Kitamura Eiji Hato discussed the probe person survey concept to obtain travel behavior data for activity- based models. He described the concept, possible approaches for a mobile activity logger (MoAL), and pilot tests of these applications in Japan. Appendix B contains a paper on this topic. The following points were covered in his presentation. ⢠Individual travel patterns vary on a day- to- day basis. Although much travel, such as commuting to work or school, is repetitive, other travel is highly variable depending on activities and needs. Travel surveys are used to obtain information on trip origins and destina- tions, time of travel, trip purpose, and mode. More extensive information on individual travel patterns is needed for activity- based models. Information on activ- ity location, activity duration, travel route, changes in travel due to external conditions, and other trip charac- teristics is of use in activity- based models. ⢠A number of survey methods are available to col- lect travel behavior information. Possible techniques include traditional questionnaire surveys and travel diaries, web- based surveys and diaries, probe person sur- veys, and surveys using multiple sensors. Each method has advantages and disadvantages, as well as cost implications. ⢠The probe person survey technique is intended to ensure accurate travel records by determining spaceâtime position using the Global Positioning System (GPS) and reducing recording omissions through timely reporting of travel behavior by cellular telephone or the Internet. This approach improves the efficiency of data coding and improves the participation of survey respondents. The probe person survey system takes advantage of recent market, social, and technology changes. First, mobile communication systems have penetrated the mar- ket. Second, cellular telephones, computers, electronic mail, and other communication technology are in wide- spread use. Third, advances in GPS, sensors, and other technologies continue to occur. ⢠The probe person survey system uses cellular tele- phones and web- based travel diaries to allow participants to easily record trips and activities. Participants can also check and correct travel records. Combining GPS and other sensors with the system allows the automatic recording of key trip characteristics. The main elements of the system include GPS- equipped cellular telephones with specially designed software, a web- based travel diary system, and a data management server. A participant acti- vates the cellular telephone at the start of a trip. Data on the location, time, and other characteristics are recorded automatically and transmitted to the server. Position data are automatically recorded every 20 seconds while the participant is traveling. A participant presses the end key on the cellular telephone upon arrival at his or her desti- nation. The characteristics of the trip are automatically recorded in a web diary system, which the participant can edit at a later time. The software for MoALs uses Brew. Oracle and Microsoft are used for the online analytical processing and the Internet information system. ⢠The MoAL system was tested in Japan over a 3- year period beginning in 2003 as part of the Matsuyama Probe Person Panel (MPPP) survey. The survey included three waves. The MPPP is a panel- type survey, with a successive diary system. Each survey period covered approximately 1 month, and new panels were added in the second and the third waves. Wave 1 of the MPPP covered 4 weeks in 2003, Wave 2 included 5 weeks in 2004, and Wave 3 covered 4 weeks in 2006. Each wave included approximately 1 million records. The partici- pant withdrawal rate was approximately 2% in Wave 1 and 9% in Wave 2. Measures were taken to protect per- sonal information. ⢠The MPPP results were compared with those of the national transportation census and person trip survey, which both cover 1 day. The MPPP results record higher numbers of trips. The average number of trips for one participant increased from 3.6 in the 2003 MPPP to 3.8 in the 2004 MPPP, while the trip omission rate declined. ⢠A limitation of the MoALs approach is that partic- ipants must enter and edit their trip activity information. Another limitation relates to the accuracy of GPS in some areas and under some conditions. A second approach, called Behavioral Context Addressable Loggers in the Shell (BCALs), attempts to address these limitations. The BCALs approach uses an automatic recording system with multisensors. The system obtains data on atmos- pheric pressure, sound, position, acceleration, and other elements. It may be possible to develop an automatic esti- mation model for behavioral contexts from these data without requiring any action on the part of participants. ⢠The results from the three waves of the MoALs pilot indicate that the use of a GPS cellular telephone and web diary system enhances the accuracy of trip reporting compared with traditional questionnaires. Location positioning can be estimated accurately with 100 meters both indoors and outside. This approach is useful for long- term detailed travel diaries. The BCALs approach holds promise as a new survey method, but privacy issues will need to be addressed. Ram Pendyala, University of South Florida, and Kay Axhausen, Swiss Federal Institute of Technology, moder- ated this session. 62 INNOVATIONS IN TRAVEL DEMAND MODELING, VOLUME 1
