Microscopic Parameter Identification of Pedestrian Models and Implications for Pedestrian Flow Modeling

Abstract

This paper focuses on the calibration of microscopic pedestrian models using pedestrian trajectory data. A generic approach that enables parameter identification for microscopic models in general and for walker models in particular is proposed. The application results provide insight into the behavior of individual pedestrians and interpedestrian differences as well as the resulting pedestrian flow characteristics. Different models of increasing complexity were compared to investigate which model amendments are significant from a statistical point of view. Besides anisotropy, finite reaction times play an important role in correctly describing microscopic walking behavior. The implications of these findings in the microscopic description of pedestrian flows are considered by studying the predicted flow operations at a narrow bottleneck. The finite reaction times have a significant effect on the pedestrian flow operations and should be included in pedestrian models to ensure realistic predictions.