Optimization of the RBF localization algorithm using Kalman filter

In this paper an optimization of the Rank Based Fingerprinting (RBF) algorithm using the Kalman filter will be introduced. In our previous papers it was shown that RBF localization algorithm seems to be more immune and accurate compared to traditional deterministic fingerprinting algorithms. We still decided to improve accuracy of the algorithm in dynamic applications, like navigation or user tracking. For this purpose Kalman filter was implemented to the previously developed localization system. Impact of the proposed solution on the accuracy and functionality of the RBF localization algorithm was investigated in the simulations and verified real world experiments. Simulations were performed in model created in Matlab environment. Real world measurements were performed at the campus of the University of Zilina using WifiLOC localization system developed at the Department of Telecommunications and Multimedia.