Hybrid methods for blind adaptive equalization: new results and comparisons

Abstract

This paper proposes two new hybrid blind algorithms based on a new radius-adjusted approach for QAM signal constellations and presents a comprehensive survey of hybrid methods for blind adaptive equalization. The proposed hybrid blind algorithms define static circular regions around symbol points that correspond to a specific weighting factor and stepsize, which optimize the equalizer tap update based on the adaptation phase. Hybrid methods are discussed for the constant modulus algorithm (CMA), improved transfer to the decision-directed (DD) algorithm, and dual-mode hybrid algorithms. Comparisons are made between the proposed algorithms and related hybrid methods, and it is shown that the new algorithms lead to enhanced performance with minimal added complexity.