Towards accurate positioning: Directionality-enhanced fingerprinting radio map

Abstract

The most essential element of location based service is accurate user localization. Considering the existing indoor infrastructure of the internet of things and WiFi access points, along with the widespread use of smartphones by individuals, fingerprinting technology emerges as the most promising indoor navigation system in terms of cost-effectiveness, performance, and accessibility. To estimate the user's location by fingerprinting, a fingerprinting radio map must be created in advance. The radio map contains received signal strength indicator (RSSI) vectors and its collection locations. In an indoor environment, RSSI values vary greatly depending on the collection direction, even if they are received from the same location. To compensate for this feature of RSSI, this paper proposes a new fingerprinting radio map that considers directionality. Since pedestrians can move in various directions indoors, all RSSI values collected along those directions are included in the radio map. In the location estimation step, the current location is estimated by comparing the spatial RSSI sequence of the current pedestrian with the reference points in different directions that have the highest correlation. To validate the proposed method, we analyzed the user location results using the existing radio map and the proposed radio map and found that the proposed radio map improved the positioning performance.