We present a method for assigning markers to locations to enable navigation by observing short sequences of markers. A motivating application is indoor navigation, where pedestrians can determine their location using the sequence of recently encountered markers as they walk along hallways in a building. While we may solve this problem simply by assigning a unique marker to each location, such a solution limits the granularity of localization due to limits on the number of distinct markers. We present a more efficient solution that uses a sequence of recently seen markers to determine locations. We demonstrate that our solution yields significant improvements over the naive solution even when the marker-sequences are quite short, consisting of only three of four markers each.
展开▼