Indoor Localization Using Smartphones in Multi Floor Environments Without Prior Calibration or Added Infrastructure

Indoor positioning for smart phone users has received a lot of attention in recent years. While many solutions have been developed, most rely on extra sensors, a need for predeployment of infrastructure or collecting ground truth data to train on. In this paper we see what can be done using only existing WiFi-infrastructure and Received Signal Strength from these, not using any extra sensors or calibration of the signal environment. We expand on previous work by using a multi floor model taking into account dampening between floors, and optimize a target function consisting of least squares residuals, to find positions for WiFis and the smartphone measurement locations. Experiments indicate that floor detection needs to be semi-supervised or in need of additional sensors. The method was tested inside two buildings, with tree stories each, with mean errors of smartphone positions of 15.2 m and 13.5 m respectively.