A Kriged Fingerprinting for Wireless Body Area Network Indoor Localization

The accurate and efficient localization of a person equipped with wireless body area network (WBAN) is still a major issue in surveillance systems. Localization mechanisms using signal strength suffer from inaccuracy but those especially based on fingerprints are likely more accurate for indoor environment and some related mechanisms have been proposed. However, a simple and accurate mechanism is still unavailable for indoor environment except those using ultra wideband technology which is not yet implemented in all existing devices. This paper proposes a new mechanism based on fingerprinting for WBAN indoor localization named Kriged Fingerprinting. From the recorded data, this proposed model generates a huge unsaved database used to estimate the mobile node. It is a hybridism of fingerprinting, particle swarm optimization (PSO) and kriging. The fingerprinting method operates into two phases: the first is the training phase that serves to build a radio map in a database and the second phase is a matching phase. PSO is used as a matching algorithm and when a particle gets to an unrecorded location, the kriging mechanism is used to interpolate its locations values. In comparison with some works done for indoor localization mechanisms and based on fingerprinting, our proposal outperforms all in terms of accuracy: the mobile node is localized with a mean error value of 0.906 m. Hence, the kriging method brings more improvement in localization based on fingerprinting.