Continuous tracking of user location in WLANs using recurrent neural networks

Location is one of the contextual variables most relevant to the design of context-aware computing systems. These applications need to know the physical location of users in order to provide information relevant to their position. Radiofrequency (RF) signals received by mobile devices can be measured to obtain the signal strength. These signals can be used to estimate the approximate location of a user. In this paper, we present a technique based on recurrent neural networks to infer user location in WLANs inside buildings. The approach uses information from previous location estimations to address the problem of continuous user tracking. This means that we take advantage of the user trajectory to reduce the inherent error causing the user to "jump" between two places separated by large distances. We present the results of the proposed approach and analysis intended to reduce the effort of measuring RF signals.