Iterative decoding of finite-state Markov channels with soft-decision
Igor Moreira, Cecilio Jose Lins Pimentel · 2017
This paper presents a non-binary finite-stat Markov channel (FSMC) that captures both the soft-decision information and the temporal correlation of a discrete fading channel (DFC) with 2ξquantization levels. The parameters of the FSMC that approximate the DFC with fixe parameters are found using the Baum-Welch algorithm. An iterative decoding scheme, based on the sum-product algorithm, is also discussed for joint decoding and channel state estimation. The performance of LDPC codes over the proposed non-binary FSMC under the joint iterative decoder is then studied and the coding gain obtained by increasing the channel output alphabet is analyzed.