An Econometric-based Model of Pedestrian Walking Behavior Implicitly Considering Strategic or Tactical Decisions

Abstract

We propose an econometric-based model for the behavior of pedestrian walking that implicitly considers strategic or tactical decisions. In real situations, it is only possible to observe pedestrian trajectories and final choices of destination (e.g., the ticket gate they have actually chosen) and the targeted destination en route is latent and unobservable. To reflect this, a model of dynamic latent plans was extensively used by assuming destination choice as a decision at the planning level and walking at the action level. The latent plans (destinations) of pedestrians may dynamically change subject to the environment, leading to a dynamical and structural change in their action choices (walking trajectories).