ABSTRACT
When placing AI-driven characters into your immersive game world, one large problem needs to be addressed, and that is the issue of a meaningful representation of the environment. The only source for information about the layout of the environment available to these characters is that which is provided to them by the game designers usually in the form of the geometric models that are assembled spatially to create the world. In all but the simplest of games, the level of detail in those model files is often too complex, too detailed, and organized more for display than spatial reasoning. Instead, some form of spatial abstraction is needed to group similar areas in single regions of space for the character to consider.
