ESTIMATING PEDESTRIANS' ROUTE CHOICE PROBABILITY IN UNDERGROUND SPACE BASED ON OD MATRIX AND FLOW COUNT DATA

Abstract

Underground spaces connected with railway stations gather a lot of pedestrians. Clarifying pedestrians' movement in underground space is important for flow control and evacuation planning. Based on a pedestrian traffic survey data and estimated OD matrix, this study aims to clarify the pedestrians' route choice probability in underground space. Through the definition of “state” and “transition process”, pedestrians' route choice model is constructed according to the utility of routes. Next, Absorbing Markov Process is adopted to solve the trip assignment in the underground pedestrian network and build the functional relation between estimated flow counts and obtained data. Through minimizing the difference between estimated and observed flow counts, unknown variables are calibrated. Finally, the influential spatial motion attributes are examined through linear regression analysis.