Modeling Collision Avoidance Field for Overtaking Moving Obstacles

Abstract

The capability of overtaking moving obstacles is an essential factor for accomplishing several aspects of pedestrian walking flow. In the Hajj area, for example, overtaking dynamic structures constituted by groups, barriers, and other moving obstacles is a vital phenomenon emerged while performing Hajj rituals. In emergency situation, the awareness of the dynamic behavior of moving obstacles is indispensable for achieving typical evacuation. This article proposes an essential intelligence approach to performing further realistic evacuation simulations. We provide each agent with the capability of selecting intermediate destination that enables him reaching his preferred destination; the agent continuously adapts his own trajectory that enables him to overtake such dynamic obstacles by selecting intermediate destinations to pass through. A collision avoidance field which composes of two-dimension grid of cells is proposed to cover the floor of the physical environment. The agent selects the optimal cells which achieve less potential of collision and minimize the distance to the original destination. The proposed model is integrated in a microscopic crowd dynamics model, and simulations are performed to examine the impact of the extended model on introducing further realistic and efficient evacuation.