Crowd Motion Control

Crowd animation has become a hot topic in resent computer graphics research. Its aspects include modeling, motion control and rendering etc. Crowd animation has applications in various situations such as evacuation design and entertainment, thus worth our efforts.Among all the aspects of crowd animation, crowd motion control seems to be the most important and valuable. In this paper we will focus on crowd navigation control. The thesis is divided into two parts:In first part we proposed a method for simulating the behaviors of large crowds based on radial basis functions (RBF). Our system adopts the vector field as a basic factor to drive the flow of a target crowd. By intuitively specifying the desired velocities on a few points in the scene, the user can easily manage the overall navigating movement of the crowd. Based on RBF inter/extrapolation, our system is fast enough to allow on-the-fly modification of the vector field. Moreover, the final simulation is leveraged by autonomous behaviors, the proportion of which can be adjusted interactively by user before or during the simulation. We have three major contributions. 1) Our method is fast and intuitive enough to allow interactive control of the crowd navigation in huge scenes. 2) Our system is flexible as it allows the user to interactively control the proportion of autonomous behavior and guided behavior during the simulation. 3) Our work (including the implemented plugin in OpenSteer) is a helpful supplement of Raynold's steering behaviors, a widely adopted technique by many animation systems both commercial and non-commercial, and OpenSteer, a widely used public steering behavior toolkit.In second part we proposed a conceptual crowd function model for an army, and proposed the corresponding implementation. This model has been verified and improved via practice (supported by the special effects making of TV series "ZhenGuanZhiZhi") and should be of help in such kind of crowd simulation in movie makings in future.