Map Learning with Indistinguishable Locations

Nearly all spatial reasoning problems in­ volve uncertainty of one sort or another. Uncertainty arises due to the inaccuracies of sensors used in measuring distances and angles. We refer to this as directional un­ certainty. Uncertainty also arises in com­ bining spatial information when one loca­ tion is mistakenly identified with another. We refer to this as recognition uncertainty. Most problems in constructing spatial rep­ resentations (map�) for the purpose of nav­ igation involve both directional and recog­ nition uncertainty. In this paper, we show that a particular class of spatial reason­ ing problems involving the construction of representations of large-scale space can be solved efficiently even in the presence of di­ rectional and recognition uncertainty. We pay particular attention to the problems that arise due to recognition uncertainty.