Abstract
The purpose of this study is to estimate the origin-destination matrix of urban area by using the activity simulation. The tour-based behavioral model which considers the interrelation of the trips was constructed by Kofu PT survey and applied it to forecasting the travel by the micro-simulation in Kofu. The estimated OD matrix by the simulation is not precise, so that we estimated the OD matrix by assimilating the observed OD matrix of Person trip survey to the simulated OD matrix. After applying the particle filter which is one of the flexible assimilation methods, the precision of the OD table was improved significantly. This study showed that data assimilation is one of the practical approaches to improve the micro-simulation of travel behavior in the big-data era.
