Point-of-Conflict Prediction for Pedestrian Path-Planning

A simulation model for pedestrian navigation often requires many problems to be addressed. An example is pedestrian path-planning in the case of a moving obstacle within the environment. To resolve the problem, it is necessary to model the prediction process of the pedestrian agent in order to specify a point-of-conflict area. In this paper, we propose an approach for our pedestrian agent to navigate in this situation. This process is accomplished by predicting a possible point-of-conflict with the obstacle and planning the path accordingly. Our implementation of this approach has demonstrated the capability of the agent to plan a more competent path as well as closer to the thinking process in human cognition.