Stochastic Transition Model for Discrete Agent Movements

TL;DR

This paper introduces a calibrated 2D cellular automaton model for pedestrian movement that accounts for direction-dependent behaviors caused by grid structure, improving simulation accuracy.

Contribution

It presents a novel 2D cellular automaton model with direction-dependent behavior and methods to compensate for grid-induced deficiencies.

Findings

Model effectively simulates pedestrian motion with direction-dependent behavior

Compensation techniques improve model accuracy

Efficiently addresses grid-caused deficiencies in simulation

Abstract

We propose a calibrated two-dimensional cellular automaton model to simulate pedestrian motion behavior. It is a v=4 (3) model with exclusion statistics and random shuffled dynamics. The underlying regular grid structure results in a direction-dependent behavior, which has in particular not been considered within previous approaches. We efficiently compensate these grid-caused deficiencies on model level.