An advanced fingerprint-based indoor localization scheme for WSNs

The purpose of this paper is to propose an advanced fingerprint-based indoor localization scheme for wireless sensor networks (WSN) to improve the accuracy. Many localization methods have been introduced for WSN systems using wireless signals mainly divided into two categories, which are range-based and range-free. As wireless ranging is not reliable in indoor environment due to multipath fading and other attenuations, range-free solutions are preferable indoors. A popular solution is RSSI fingerprint-based algorithm or its variances within KNN, R-KNN, and WKNN, which can achieve accurate localization without interference. However, random and unpredictable human presence and movement cause certain level of interference to reduce the accuracy of indoor localization. In this paper, LWMA scheme is introduced to filter interfered RSSI values and contribute to algorithm recover back to no interference condition. The experiments show that the scheme can improve the average positioning accuracy by 50% with standard deviation decrease 40% in interfered condition.