Adaptive Constant Modulus Blind Equalization with Variable Step-Size Control by Error Power

A variable step-size control by error power was proposed to improve the performance of adaptive constant modulus blind equalization algorithm based on the analysis of the cost function. In the Constant Modulus Algorithm (CMA), large step-size can achieve faster convergence rate, but the steady-state residual error is big, on the contrary, small step-size can achieve higher convergence precision, but the convergence rate is slow. A triangle inequality can be set up according to the cost function of CMA and the power of error and signal can be estimated by exponential decay window, then the attenuation function can be obtained to control the step-size change. Meanwhile, the threshold is set to reset the step-size to ensure the tracking performance when the channel has burst interference. Computer simulation results prove the effectiveness of the proposed algorithm.