Abstract
This study argues the activity based model (ABM) simulation can increase its traffic demand predictability by assimilating the Population Distribution estimated from Mobile Phone network data (PDMP) on the central district of Tokyo. ABM has developed to forecast the travel demand in the last twenty years. It, however, faces still some challenges over its validity in forecasting characteristics of activities. Meanwhile, Mobile Network Operators (MNO) has started providing the population estimate on census meshes based on PDMP. MNO presents the observation of the time-sliced population in zones, while the ABM is to forecast the popular traffic demand for certain period. This study will assimilate AMB and PDMP to improve the validity of traffic demand forecast. We found that such assimilation did improve the usefulness of the ABM.
