Simulating virtual crowd with fuzzy logics and motion planning for shape template

Abstract

Simulation of crowd motions has great potential in many applications in robotics, games and animations. However, it is also a great challenge to be able to control the motion of a virtual crowd according to the intents of its designer such as making the crowd conform to a specific shape while avoiding collisions with other agents in the crowd or with obstacles. In this paper, we propose a simulation mechanism that works in two steps: global motion planner for a shape template and fuzzy controller for shape constraints. The system first uses a motion planer to generate a rough path for a desired shape that may be partially in collision with the environment. Then a fuzzy controller based on several criteria is used to move the agents in a group to conform to the desired shape. We will demonstrate the implemented system with several simulation examples that show the path of the shape template and how the crowd effectively conforms to the template.