Least square kurtosis constant modulus algorithm based underwater acoustic channel blind equalizer

Constant modulus algorithm (CMA) blind equalizer has advantage of computing efficiency and disadvantages of slow rate of convergence and large residual mean square error (MSE). Least mean kurtosis CMA (LMKCMA) has to estimate the MSE, and this estimation has a bad influence on its expectation behavior. For greatly overcoming these disadvantages, a new cost function based on kurtosis of error signals is defined and analyzed, a least square kurtosis constant modulus algorithm (LSKCMA) for updating weight vectors of blind equalizer is proposed. In the LSKCMA, the kurtosis factor based on error signals can improve convergence rate and make the algorithm converge to global minima. Thus, the LSKCMA has much faster convergence rate than the LMKCMA and the CMA. Simulation results with negative acoustic gradient underwater channel equalization have shown the efficiency of the LSKCMA.