ABSTRACT

Now that we’ve covered the four key components for success, let’s examine the models themselves.

Any game where the player can hide, or break line of sight (LOS), requires some type of visual perception model. Commonly, we refer to this as a vision cone, and while a cone does a good job of modeling what the NPC can see directly in front of him, it does a poor job of modeling many other aspects of vision. The two most glaring examples are peripheral vision and vision at a distance. It’s pretty obvious why a cone doesn’t model peripheral vision well, but for vision at a distance, we need to dive a bit deeper.