Indoor positioning based on improved weighted KNN for energy management in smart buildings

Abstract

Offering special service to residents of smart buildings to achieve the energy efficiency entails knowledge of identity information, place of residence and also the current location of people inside the building. However, localization accuracy adversely degrades in non-line-of-sight (NLOS) environments. In this study, we design a low-cost indoor positioning system based on the Wi-Fi fingerprint embedded on the smartphones. Indoor positioning system is composed of two online and offline sections. In the offline phase, a platform for collecting the radio map information is introduced. Then, the noise covariance of the received signals is estimated by adaptive Kalman filter. In the online phase, online layer clustering and K-nearest neighbor method based on the fisher information weighting and differential coordinates are presented. Simulation results show that the proposed method improves errors of less than 2 m by 40% compared to other methods. Also, the proposed algorithm is comparable to other algorithms in terms of computational complexity.