An hidden Markov model for indoor channel based on experimental data

Accurate models of digital transmission channels is fundamental in the assessment of the performance of communications systems. In particular generative models for their usability in simulations have been studied through the years. In this paper we describe a model using deterministic hidden Markov systems customized to match experimental data and provide some rule of thumb to drive the architectural choices of the state diagram. Model parameters are tuned by means of the Baum-Welch algorithm starting from measured error patterns in an indoor environment in the 1.9 GHz band for different receiver speeds and interference levels.