Flocking Boids with Geometric Vision, Perception and Recognition

In the natural world, we see endless examples of the behavior known as flocking. From the perspective of graphics simulation, the mechanics of flocking has been reduced down to a few basic behavioral rules. Reynolds coined the term Boid to refer to any simulated flocking, and simulated flocks by collision avoidance, velocity matching, and flock centering. Though these rules have been given other names by various researchers, implementing them into a simulation generally yields good flocking behavior. Most implementations of flocking use a forward looking visual model in which the boids sees everything around it. Our work creates a more realistic model of avian vision by including the possibility of a variety of geometric vision ranges and simple visual recognition based on what boids can see. In addition, a perception algorithm has been implemented which can determine similarity between any two boids. This makes it possible to simulate different boids simultaneously. Results of our simulations are summarized