ABSTRACT
In the laboratory experiment, route choice behaviour was analysed as a basic study for discussing the possibility of route dispersion via navigation through information provision. The experiment was aimed at clarifying driver's route choice behaviour under the provision of information on travel time. The actual travel time, required for a trip on a route chosen by the subject, was presented to make him/her recognize the travel result. The relationship between difference in travel time according to information and the choice ratio of route 1 was analysed to determine the influence on driver's route choice of difference in the accuracy of provided information. In estimating parameters of the model for travel time prediction, it was therefore necessary to use the generalized least-square method enabling the consideration of the serial correlation of errors.
