Effectiveness of online RF fingerprinting for indoor localization

Abstract

This paper proposes a method for indoor localization by introducing an online Radio Frequency (RF) fingerprinting approach instead of the traditional offline RF fingerprinting. The offline phase is a time-consuming stage where manual measurements made by the user cannot be at high accuracy due to the fact of shadowing effect that human body causes during the construction of RF radio map. Moreover, outdated RF radio map is another issue when it comes to the dynamic environment conditions (displacement of objects, moving people and/or the opening and closing of room doors). By proposing online fingerprinting, the problems mentioned above are solved by obtaining real-time RF maps and estimating an unknown location accordingly. Our method is unique when it comes to the approach, devices and the environment of deployment, and it has successfully shown up to 23% improvement in localization accuracy