Analysis of passengers' route-choice behavior in vertical transportation by support vector machine

The simulation of pedestrian movement is important to the architectural design of a building to ensure the effective movement in crowd and emergency conditions. Route choice is one of the decision making process in the pedestrian movement. In current practice, scientists and engineers obtain the probabilities of route choice from the statistical results of the experimental data. The probabilities are also assumed to be constant. In reality, the routechoice behavior of a pedestrian depends on many factors. We adopted support vector machine in this study to simulate the decision made by a pedestrian in choosing either staircase or escalator for vertical transportation based on the environmental conditions visually observed by a pedestrian. The model achieves 90% of correct prediction.