Data processing algorithm for direction tracking in indoor localization

Indoor localization based on inertial sensors such as accelerometer, gyroscope and magnetometer, has gathered broad interest within the scientific community. This approach in fact does not require the installation of external devices since it relies on low-cost devices of common use, like smartphones, which are used both as measurement platform and user interface. In this paper we focus the attention on implementing algorithms for estimating and tracking the heading direction of a user while he/she is walking inside a building. The novelty of the method here proposed, is that it does not refer to an absolute estimation of the heading direction of a user with respect to the four cardinal directions, as usually done in literature. The method proposed in this paper is based on detecting relative changes in the direction of a user with respect to a reference system obtained during an initial calibration phase. Then, the fundamental direction which exhibits the minimum ℓ₂-norm distance with respect to the raw measured values is provided in output. As it can be noted from experimental results reported in the paper, the direction of a user can be detected with a sufficient reliability.