Research on fingerprint localization algorithm based on multivariate Gaussian mixture model

To settle the matter which high-dimensional classification problem caused by the large number of access points and the varying levels of signal reception intensity recorded by heterogeneous terminals, a multivariate Gaussian mixture model is used to simulated the portion of received signal strength (RSS) of each unit with effect, and a Wi-Fi fingerprint algorithm based on hidden Markov model is used to track moving users. Finally, the accuracy of the algorithm was verified through user crowdsourcing data. According to the experimental results, that the algorithm achieved a reliable matching accuracy of 92.98%, increasing from an average matching accuracy of 92.98% to an overall matching accuracy of 97.29%.