Tensor-based blind identification of MIMO Volterra channels in a multiuser CDMA environment

This paper is concerned with the blind identification of Multiple-Input-Multiple-Output (MIMO) Volterra channels in a multiuser Code Division Multiple Access (CDMA) environment. The channel is modeled using the most generic representation of complex-valued Volterra systems. A Parallel Factor (PARAFAC) decomposition of a third-order tensor composed of channel output covariances is used, the transmitted signals being assumed to be Phase Shift Keying (PSK) modulated. The channel estimation is carried out by two algorithms: the Alternating Least Squares (ALS) algorithm and a non-iterative least squares algorithm that exploits the redundancy provided by the Khatri-Rao product. The performance of the proposed estimation methods is illustrated by means of computer simulations.