GPS/GIS localization for management of vision referenced navigation in urban environments

This communication presents localization in urban environments as a mean to construct and retrieve visual landmarks for vision based navigation. It deals with localization in a navigation map of the urban environment which is known to be inaccurate. A fuzzy sets formulation is shown to be convenient and localization is performed by computing ellipsoidal outer envelops of /spl alpha/-cuts of the pose estimation. Fusion between data from hybrid GPS and the navigation map is performed by using ellipsoidal algorithms of set membership estimation.