Abstract
This paper proposed an hourly OD flow estimation model from observed traffic flows based on time coefficients of daily OD flow and examined the accuracy of estimation through the application to a large-scale road network. Firstly, it was cleared that there is some bias in the time coefficients of daily OD flow aggregated from questionnaire survey. The bias means that time coefficients of daily OD flow are overestimated in peak hours and underestimated in night-time, especially for trucks. Therefore, we developed the time coefficients estimation model from observed traffic flow by using the least squares estimator under a given daily OD flow through the application of time-of-day user equilibrium traffic assignments. It was demonstrated that the model can improve the accuracy of time-of-day traffic flow estimation as compared with the traffic assignment using an original hourly OD flow. And the equation transforming the estimated hourly OD flow into OD flow based on departure time was also shown.
