Author(s)

Y. Yamakawa & E. Hato

Abstract

One of the difficulties of route choicemodeling is to represent a generation process of choice set explicitly in a network including infinite paths. Since it is expensive and uses many man-hours to collect the data, the analysis of routes that a driver actually used has not been done. Therefore, in the existing route choice model, there is a problem that routes included in a choice set do not necessarily correspond to routes that a driver has actually used. However, with the development of Probe Person Technology that gives location data of the movable body, having a GPS phone, it is possible to collect the data of routes a driver has used without burdening the driver. Then, in this paper, the route data which is available on a private basis and obtained in Probe Person survey is analyzed. The choice set generation methods based on routes drivers actually used are suggested, and the case studies of route choice models are carried out by using it. keywords: route choice modeling, choice set generation, GPS data. 1 Introduction The route choice model is used for the traffic simulation and is vital to the analysis of road networks. But, the applicability of route choice models for traffic simulation is not sufficient, it is attributed to how to generate route choice set. When we choose our routes in the road network, we must restrict the road network and generate the choice set in some stages (Bovy [1]). The following four processes are used to generate the choice set, Existing set→Known set Available set→Feasible set→Choice set. Though it is arguable that we choose our routes through these steps, actually, capturing processes to generate the choice set explicitly is essential for the route choice modeling. The Analysis of routes drivers actually used has not been made until now because it is expensive in direct costs and

Keywords

route choice modeling, choice set generation, GPS data.