Estimação de Parâmetros de Canais de Gilbert-Elliott

This paper compares three methods for estimating the parameters of class of finite state Markov channels, specifically the Gilbert Elliot channel (GEC), so as to approximate a discrete fading channel. For this, the method proposed by Wilhelmsson and Milstein and the Baum-Welch algorithm, already used in the literature, are compared with the minimization of the Kullback-Leibler divergence between two probability measures. The accuracy of the GEC obtained with each estimation technique is measured in terms of the autocorrelation function, the throughput of the Go-Back-N protocol and the probability of an unsuccessfully decoding of a block code.