An EM based Joint Data Detection and Channel Estimation Incorporating with Initial Channel Estimate

In this letter, we derive an expectation-maximization (EM) algorithm to solve the maximum-likelihood (ML) detection problem over multiple input multiple output (MIMO) channels. The difference from the existing EM algorithms for the ML detection problem is that the proposed EM algorithm does not require the probability density function (pdf) of channel parameters. Thus, the proposed EM algorithm is applicable for any type of MIMO channels (e.g., Rician or Rayleigh) without any modification.