The paper deals with the problem of blind source extraction from a MIMO convolutive mixture. We define a new criterion for source extraction which uses higher-order contrast functions based on so called reference signals. It generalizes existing reference-based contrasts. In order to optimize the new criterion, we propose a general algebraic algorithm based on best rank-1 tensor approximation. Computer simulations illustrate the good behavior and the interest of our algorithm in comparison with other approaches.

No abstract

The paper deals with the problem of the blind extraction of a source signal after a MIMO convolutive mixture. The extraction is performed using a MISO equalizer. A contrast function based on high order statistics is first proposed. It is more general than the existing contrast in the same context and exhibits a cubic dependence w.r.t. the unknown equalizer parameters. This allows us to propose a new extraction algorithm based on a third order tensor decomposition. Computer simulations illustrate the good behavior and the usefulness of our algorithm.

No abstract

No abstract