An adaptive hybrid filter for practical WiFi-based positioning systems

This paper proposes an adaptive hybrid filter for WiFi-based indoor positioning systems. The hybrid filter adopts the notion of particle filters within the prediction framework of the basic Kalman filter. Restricting the predicts of a moving object to a small number of particles on a way network, and replacing the Kalman gain with a dynamic weighting scheme are the key features of the hybrid filter. The adaptive hybrid filter significantly outperformed the basic Kalman filter, and a particle filter in the performance evaluation at three test places: a Library and N5 building, KAIST, Daejeon, and an E-mart mall, Seoul.