Simulating Crowd Motion with Shape Preference and Fuzzy Rules
Jen-Yao Chang
Crowd simulation is becoming indispensable in computer games and animations. Several approaches to this problem have succeeded in generating flocking behaviors with virtual forces but it remains difficult to manipulate the collective behaviors of a crowd to respect certain desired shape. In this paper, we propose a crowd simulation system that can generate a feasible path taking the crowd to the goal while maintaining a user-specified shape as much as possible. The system consists of two key components: a path planner that can search for a feasible path for a region of flexible shape and a local motion controller based on three classes of fuzzy rules for generating desired agent behaviors. We will demonstrate the implemented system with several examples to show the generated path and the corresponding crowd motions adhering to a desired shape.
Category: Virtual Crowd Simulation