Semi-Blind Stochastic Maximum Likelihood for Frequency Selective MIMO Channels

We propose a semi-blind stochastic maximum likelihood (SML) estimator for frequency selective multi-input multi-output (MIMO) channel. To resolve the optimization problem we use the expectation-maximization (EM) algorithm. We first propose a hidden Markov model (HMM) for the MIMO frequency selective channel which gives us a formulation of the EM algorithm by introducing forward and backward parameters. Previously, semi-blind SML methods were provided for time-multiplexed pilot and data symbols. In this work, the semi-blind SML is proposed for the embedded pilot scheme which is attractive for its spectral efficiency. We also compare the semi-blind SML for the embedded pilot scheme with the blind SML and show the benefit of using semi-blind SML over blind SML