Computationally Eecient Blind Equalization

Adaptive blind equalization has gained widespread use in communication receivers that operate without training signals. In particular, the Constant Modulus Algorithm (CMA) 6, 10] has become a favorite of practitioners due to its LMS-like complexity and desirable robustness properties 4, 5, 14]. This paper presents a signed-error version of CMA (SE-CMA) which is motivated by a further reduction in the computation complexity of the equalizer coeecient update. The authors investigate the performance of SE-CMA and establish connections to previously proposed signed blind algorithms. A novel dithered signed blind algorithm (DSE-CMA) is proposed which overcomes signiicant SE-CMA convergence issues and possesses similar robustness properties to the standard (i.e. unsigned) CMA.