Using blind source separation algorithm by minimizing mutual information in reference channels of multichannel adaptive cancelling disturbances

 

A. E. Manokhin

Ural Federal University

 

 The paper is received on December 3, 2015 after correction - on March 7, 2016

 

Abstract. In the work the problem of multichannel adaptive cancelling disturbances with their preliminary separation in reference channels is decided. A kind of the canceller is an application of blind separation algorithm by minimizing mutual information. As a result, first, convergence rate is increased, and second, disturbances separate even if to be exposed by different linear transformations in a channel. The gain of convergence rate equals not less than 28 times.

Key words: blind sources separation, multichannel adaptive canceller, algorithm by minimizing mutual information, convergence rate.