Floor level estimation using MEMS pressure sensors

Abstract

Enhanced 911 systems aim to increase the efficiency of first responders by automatically locating the origin of a 911 call even if the caller is unable to provide it. However, in a high-rise building, the location of a cell phone call would appear the same to dispatchers regardless of elevation. If emergency services had more accurate location information, they could save countless more lives. In this study, we compared multiple methods for determining a caller’s floor level inside a building, including GPS, WiFi, a magnetic sensor, and a micro-electromechanical system (MEMS) pressure sensor. GPS provided a reasonable estimate of vertical height, but accuracy was strongly dependent on the reception, which was often poor inside the building. WiFi signals and magnetic fields were not strong predictors of vertical height. WiFi was limited by short signal range, need for network access, and knowledge of router locations. Magnetic sensor readings required a pre-existing magnetic field contour map of the building in order to interpret the data. The MEMS pressure sensor was the most accurate predictor of elevation and does not rely on communication with other equipment. A detection accuracy of ±1 floor can be achieved using the MEMS pressure sensor after correcting the output using publicly available data and assuming an average floor height. The proposed technique provides a self-contained solution, using a sensor available in most smartphones, to determine the floor of a caller. Adding floor level data to emergency call locations would provide valuable information to first responders in densely populated urban areas.