A probabilistic indoor localization algorithm based on Restricted Boltzmann Machine

With the fast development of Location Based Service (LBS) applications in recent years, the demands for accurate indoor localization techniques have attracted significant attention and risen rapidly. Among all the techniques, due to its stable performance without the need for additional hardware, the RSS-based fingerprinting localization is the most viable method. However, the traditional methods do not make full use of the energy-based model, which actually affects the accuracy of positioning a lot. In this paper, an improved probabilistic localization algorithm named Weighted Restricted Boltzmann Machine (WRBM) is proposed, which takes the energy-based model into consideration. By calculating the related probabilistic function, the proposed algorithm gets higher accuracy. As shown in the experimental results, the proposed algorithm performs much better than the other typical fingerprinting localization methods.