Smart probabilistic fingerprinting for WiFi-based indoor positioning with mobile devices

Different positioning schemes are based on the probability p(o|l) to have an observation vector o at a Reference Point (RP) l, based on Gaussian probabilities. This paper presents an approach to speed-up the p(o|l) computation without any approximation. The consequent positioning scheme is called Smart P-FP. The comparison between Traditional (without any p(o|l) acceleration) and Smart P-FP is performed over different smartphones. The saved energy is about 90% for a large number of Access Points (APs) but is significant even with few APs: more than 86% with 3 APs. The proposed p(o|l) computation is beneficial to any p(o|l)-based positioning scheme.