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.
monetary expenditures and types of commodities and services purchased at commercial establishments. These kinds of data will make survey results applicable to a wider range of planning studies. This paper summarizes the results of several pilot sur- veys conducted in Japan by using probe person systems. It shows technical requirements for new travel behavior surveys that are based on the results of these pilot sur- veys and discusses the possibility of data- oriented travel behavior analysis. TECHNICAL REQUIREMENTS FOR TRAVEL BEHAVIOR SURVEYS An individualâs activity and travel vary from day to day. To capture day- to- day variations in travel behavior in urban space through surveys and to develop and evalu- ate alternative transportation policy measures, one would desire to 1. Implement a long- term survey, 2. Determine spaceâtime coordinates with high reso- lution and accuracy, and 3. Automatically and accurately measure detailed activity contents. First, to capture more accurately the diverse patterns of travel behavior that vary day to day, it is necessary to implement a long- term, detailed travel behavior survey and observe day- to- day variations in travel patterns. A one- day survey would be sufficient if one wishes to acquire information on highly repetitive travel such as commuting. Travel patterns vary from day to day for various reasons, as noted above, and oneâs knowing the patterns of variation is a prerequisite for determining the causes for, and revealing the mechanisms of, day- to- day variations in travel. This knowledge obviously requires long- term observation of travel patterns. Second, it is desirable to determine the spaceâtime coordinates of an individualâs trajectory in urban space along the time axis. Gathering this information calls for identifyingâ with high accuracyâ activity locations, activity durations, trip starting and ending times, and the like. In some planning contexts, one would also desire to determine travel routes, transfer locations, public transit waiting times, and the beginning and ending times and locations of the respective trip segments typically involved in a public transit trip. (Even an automobile trip involves access and egress walking, yet the exact location of parking is often ignored in conventional travel sur- veys.) A survey method aimed at capturing the total daily demand volume between zones is incapable of providing the information necessary to evaluate planning measures intended to manage travel demand by fine- tuning a range of factors that influence service levels. A survey method that provides timeâspace coordinates of an individualâs movement trajectory with high resolution is desired. Third, there is a fundamental issue of whether trip measurements obtained from conventional travel sur- veys may be inadequate to represent an individualâs travel pattern in time and space. For example, accurate measurement of travel behavior calls for the recording of short trips. Yet, in addition to the fact that conventional questionnaire surveys have a problem with reporting omissions, it is often difficult or costly to geocode trips accurately on the basis of the information available from conventional questionnaire surveys, and errors can be excessive in the case of short trips. Moreover, the accu- racy of measurements that are based on survey respon- dentsâ memory may be inadequately low in relation to the current planning analysis requirements. IMPLEMENTATION OF SURVEY TECHNIQUES This section covers the characteristics and problems of the following five survey methods as tools to collect data for travel behavior analysis: 1. Questionnaire surveys, 2. Web surveys, 3. Mobile phone surveys, 4. Probe- person surveys, and 5. Surveys using multiple sensors. A questionnaire survey can be administered in many ways, each having advantages and disadvantages. For example, it is relatively easy to collect data at a low cost per completed survey but with a low response rate and a lower quality of data in the case of mail- outâmail- back surveys. These problems may be overcome when a sur- vey involves home visits (either just to deliverâcollect questionnaires or to conduct home interviews) but with the much higher costs and difficulties of visiting house- holds. Moreover, either method requires data coding, which consumes monetary cost and time in proportion to the sample size. Computer- aided telephone interview (CATI) surveys have become a standard procedure in household travel surveys in the United States. It has many advantages (e.g., data coding is automated, branching is automatic and error free, and questionnaires can be easily cus- tomized) as well as disadvantages (e.g., it takes an exces- sive amount of interview time in case of a large household). As with conventional surveys, trip records are obtained on the basis of the respondentsâ abilities to recall travel events on the survey day. The quality of the data also rests on the respondentsâ understanding of what exactly a trip is and how its attributes, including 188 INNOVATIONS IN TRAVEL DEMAND MODELING, VOLUME 2