Modeling the received signal strength intensity of Wi-Fi signal using Hidden Markov Models

Wi-Fi fingerprinting is one of the methods that are widely used to provide Location Based Services (LBS). Gaussian, or a mixture Gaussians, preferred model by for LBS. Nevertheless, Received Signal Strength Intensity (RSSI) histograms skewed, and Gaussian not well suited modeling data when their histogram skewed. In addition, another important characteristic present in RSSI temporal series autocorrelation, which cannot be modeled using model. this paper, we explore feasibility Hidden Markov Models (HMM) signals. The mathematical derivation formulas calculate autocorrelation based on HMM parameters presented. Exhaustive experimentation, sampled real scenario, was performed test dependency coefficients number hidden states, iterations creating HMM. results compared with calculated data. Kullback–Leibler (KL) divergence compare similarity those provided Gaussians an models reported more accurate than both cases.