Indoor Positioning and Pre-processing of RSS Measurements

Rapid expansions of new location-based services signify the need for finding accurate localization techniques for indoor environments. among different techniques, rss-based schemes and in particular one of its variants which is based on graph-based semi-supervised learning (g-ssl) are widely-used approaches the superiority of this scheme is that it has low setup/training cost and at the same time it leads to low localization error. analyzing the g-ssl method we can observe that its performance is highly dependent on its inputs (rss measurements). the main objective of this work is to further improve the accuracy of g -ssl based schemes by performing multiple rss measurements and then passing them through pre-processing blocks to improve the reliability of the corresponding rss vector at each sample points (sps). experimental results are then followed to show the performance of the proposed method compared to what we get with the original g-ssl approach.