Sensor fusion for floor detection

Floor identification is an important aspect of indoor positioning in multi-story buildings. The ubiquity of wireless access points (e.g. WiFi) and the integration of micro electro-mechanical sensors, e.g. barometers in mobile handheld devices during the past few years, have motivated the research and development efforts in this area. Received signal strength (RSS) and barometric pressure (BP) sensing methods can provide coarse floor identification without additional infrastructure. However, the low accuracy afforded by RSS-based methods (~22% error rate) and the susceptibility of stand alone BP techniques to environmental elements, have limited suitable applications. In this paper, a novel floor detection algorithm is developed and tested based on fusing BP and RSS measurements using Kalman Filter. As demonstrated by our experimental results, our proposed method has a staggeringly low error rate of ~0. 5% The proposed technique does not require new infrastructure, hence, can be readily integrated into mobile devices of current vintage.