In this study, taking the pathways conflict situations as a condition to construct a model of pedestrian movement behavior, by marking coordinates of conflict pedestrian and the people who around the pedestrian at each time and place, using the concept of the lattice, the range of the information to import information and analyzed in order to understand the moving behavior of pedestrians. However, under the same conditions of information, this study research different processing platforms, including rectangular platform and fan-shaped platform, which setting concepts are different, including dividing the number of space variables and setting pedestrian territorial range, then using SPSS software for multiple logistic regression model calibration by analyzing Goodness-of-fit and Pseudo R-Square. Research result is taking rectangular platform as the processing platform is more appropriate. In this pedestrian movement behavior, when a conflict occurs, the responsive area is contained in the endless belt shape area which about 0.97 meters. The distance between pedestrian about 58.21 cm to 136.17 cm, the distance and pedestrian walking speed in this area does not have significant correlation.
In this study, the model used to predict the pedestrian movement behavior, the predicted rate of 78%. If you dismiss the special walking behavior due to psychological factors of the pedestrian, then use this model to predict pedestrian movement behavior does have its feasibility.