SPATIAL REPRESENTATION FOR NAVIGATION IN ANIMATS

This article considers the problem of spatial representation for anima t navigation systems. it is proposed that the global navigation task, or ''wayfinding,'' is best supported by multiple interacting subsystem s, each of which builds its own partial representation of relevant wor ld knowledge. Evidence from the study of animal navigation is reviewed to demonstrate that similar principles underlie the wayfinding behavi or of animals, including humans. A simulated wayfinding system is desc ribed that embodies and illustrates several of the themes identified w ith animat navigation. This system constructs a network of partial mod els of the quantitative spatial relations between groups of salient la ndmarks. Navigation tasks are solved by propagating egocentric view in formation through this network.