An improved CMA-based hybrid algorithm for blind channel equalization

Abstract

Blind channel equalization has gained great importance in the world of communications. Among a large number of available blind equalization algorithms, the CMA (constant modulus algorithm) enjoys widespread popularity because of its LMS-like complexity and robustness However, as the need for faster blind equalization of a variety of physical channels (including mobile and wireless ones) increases, the search for more efficient and faster algorithms becomes paramount. This paper proposes one such hybrid CMA-based algorithm that exploits the advantages of the well-known concepts of signed-error update, pre-whitening and dithering. As such, it enjoys a simpler structure and faster convergence than other CMA-based algorithms, while retaining the robustness of the original CMA algorithm. The extensive simulation carried out here corroborates very well the claim that the proposed algorithm outperforms various existing algorithms, including the DSE-CMA and PW-CMA. An attractive feature of the proposed hybrid algorithm is that it is particularly suited for channels where ill-convergence needs to be treated with minimum additional complexity and without any loss of robustness.