Identifiability conditions for blind and semi-blind multiuser multichannel identification

We explore the identifiability conditions for blind and semi-blind multiuser multichannel identification. Starting with the deterministic approach, we compare the identifiability conditions on the channel and the sources of training sequence-based channel identification, blind and semi-blind channel identification. Further on, we use the stochastic approach with Gaussian priors for the source symbols and do the same derivations. Comparison between the two approaches lead to the conclusion that the latter yields less restricting conditions and, moreover, that the conditions are sufficient for any stochastic approach