High Fidelity Visual Representation of Crowds

Soldiers in remote territories need to be able to quickly assess the prevailing threat level from crowds in urban settings. Existing training solutions, however, do not generally provide realistic crowd simulation capabilities, as commodity graphics hardware has not, until recently, been up to the task. To generate effective crowds in a typical visualization environment involves the generation of hundreds of characters, which may exert a heavy load on graphics systems already taxed by terrain, structures, vehicles, and various scene clutter. In addition, plausible crowd behavior, both for individuals and the collective, is an equally important element in realistically simulating crowds. These behaviors must be scriptable and dynamically controllable to provide the flexibility necessary to accurately prepare troops for any urban situations they may encounter. The aim of this project is to design and develop a complete crowds framework to populate a virtual environment with hundreds of characters. Crowd behavior will be governed by a cognitive model designed to allow the flexibility to provide high level scripting, while simultaneously allowing for the collective emergent behaviors so fundamental to real crowds.