An advanced algorithm for Fingerprint Localization based on Kalman Filter

When it comes to Fingerprint Localization, the quality of fingerprint database is important to the accuracy of the localization results. As the received signal strength(RSS) collected in offline phase for Fingerprint Localization is usually companied with noise, it severely degrades the accuracy of the final localization results. To filter the noise, this paper proposed an advanced algorithm based on Kalman Filter (KF). The algorithm at first uses KF to filter noise of the measured RSS. Then it chooses several calibration points according to the weight of filtered data. Last the position of the user has been estimated according to the location of these calibration points. The experiment results illustrated that the algorithm proposed in this paper had improved the accuracy of location estimation efficiently.