GIPSy: Geomagnetic indoor positioning system for smartphones

A low complexity and infrastructure free indoor positioning system is proposed, by tracking the users on a set of user profiles via Viterbi algorithm. A modification on the implementation is also proposed to further optimize the complexity of tracking, by using the sparsity of the state transition probability matrix. Our algorithm only requires users to hold a smartphone, in arbitrary orientation/gesture and walk in the building. Since we only use the sensors embedded on the smartphones, this approach does not require any installed infrastructure in the building. The algorithm is tested on real-case scenarios with successfully obtaining less than 2.5 meters of median error.