Indoor Positioning System using Walking Pattern Classification

In the age of automation the ability to navigate persons and devices in indoor environments has become increasingly important for a rising number of applications. While Global Positioning System can be considered a mature technology for outdoor localization, there is no off-the-shelf solution for indoor tracking. In this contribution, an infrastructure-less Indoor Positioning System based on walking feature detection is presented. The proposed system relies on the differences characterizing different human actions (e.g., walking, ascending or descending stairs, taking the elevator). The motion features are extracted in time domain by exploiting data provided by a 9DoF Inertial Measurement Unit. The positioning algorithm is based on walking distance and heading estimation. Step count and step length are used to determine the walking distance, while the heading is computed by quaternions. An experimental setup has been developed. The collected results show that system guarantee room level accuracy during long trials.