Fast converging semi-blind equalization algorithm for QAM signals

This paper focuses on the problem of intercrossing interference of signals and intersymbol interference for dispersive multiple-input multiple-output (MIMO) systems employing quadrature amplitude modulation (QAM) signals. An improved soft decision-directed (ISDD) algorithm is proposed. The proposed algorithm accurately utilizes constellation information of QAM signals and avoids the misadjustment presented in classical constant modulus algorithm and multimodulus algorithm, which results in an improvement of the equalization performance. Furthermore, the cost function is optimized by gradient-Newton method. The proposed algorithm converges much faster than the gradient-type algorithms, which are usually used at present. Mean square error and maximum channel distortion are used to analyze the reliability and validity of the proposed algorithm, respectively.