Behavioral animation model for real-time crowd simulation

We present a generic model which can handle different aspects of crowd simulation and provide the animator with the ability to author different scenarios. In our model, we adopt multi-level behavior hierarchy based on rules, finite state machines (FSM) for controlling groups' or agents' behavior, which can provide simple crowd behaviors, as well as complex ones. At the same time, behaviors can be both programmed and autonomous, also composed on the fly for enabling agents respond to changing environment appropriately. We demonstrate our model through manifest the crowd simulation including the predator and prey flocks.