A unidirectional agent based pedestrian microscopic model

The goal of this paper is to establish a pedestrian model that is capable of simulating pedestrian movement at the microscopic level under normal conditions. Such a model would be useful in studying pedestrian-related transportation issues that require predicting accurate trajectories of pedestrians. The model utilizes the agent based modeling approach in which a continuous space was used to represent the pedestrian environment. This paper presents the first step of the model development: details of a unidirectional micro-simulation pedestrian model along with its calibration and validation. The model was validated by comparing simulated trajectories with real trajectories extracted from video footage by means of computer vision techniques. Results show that the average location difference between simulated and actual trajectories was 20 cm in the X direction and 40 cm in the Y direction. Results also show that the model is capable of predicting pedestrian speed with relative error of 12.8%.