Modeling detour decision combined the tactical and operational layer based on perceived density

Abstract

A microscopic model combining Voronoi-based detour SFM of the operational level and vision-based path decisions of the tactical level is developed in order to improve pedestrian detour behavior simulation at the operational level. The model is used to simulate the detour behavior of pedestrians in the circle antipode experiment. By modeling avoid crowd decision process based on visual perception, and take the path direction as the target direction of the Voronoi-based detour SFM, this model can well describe the detour behavior of the circle antipode experiment. The central area density, the mean local density, the average speed time series and the travel time, detour level, route length distribution are chosen as indexes, and then the Kolmogorov–Smirnov (K–S) test and the Dynamic Time Warping (DTW) method are applied to evaluate these indexes of the traditional SFM, Voronoi-based detour SFM and this Double-Layer Decision-Making model. Results show that the indexes of the Double-Layer pedestrian simulation model are superior to those of other models and significantly superior at the detour level, the travel time, and the route length.