Comparison of floor detection approaches for suburban area

Rafał Górak , Marcin Luckner

Abstract

As a part of smart-buildings, indoor localisation systems – alternative to Global Positioning System localisation – bring constantly improving results. Several localisation methods works with a horizontal localisation error less than few meters. However, for small suburban houses, horizontal localisation is not as important as detection of the current floor, which in is still a challenge in multi-storey buildings. This paper compares several approaches that can be used in fingerprintingbased floor detection systems. The tests include the following fingerprints: pressure measures, Wi-Fi signals, and two generations of cellular networks signals. The tests have been done in the suburban 3-storey building with underdeveloped Wi-Fi and cellular infrastructure. Notwithstanding, the floor detection based on Received Signal Strength from both infrastructures reached from 98 to 100%. Additionally, we showed that differences in the number of measures and differences in the number of received signals were not a major factor that influenced on accuracy.