Probabilistic step and turn detection in indoor localization

. In this paper we present a method to estimate a position in buildings without using absolute positioning data like WiFi signals. Unlike other state-of-the-art methods, we use probability estimations for possible steps and 90° turns. The only information source we use is the data collected by a smartphone's accelerometer and gyroscope and the floor map information. The current position is tracked with the help of particle filtering. For this, we integrate the information of the previous state into the weight update step. In addition we show how the observation data can help within the state transition model.