Variable Step Size Blind Equalization based on Sign Gradient Algorithm

Abstract

This work proposes a variable step size sign gradient algorithm to solve the problem of blind equalization under impulse noise environment. This algorithm suppresses the impulse noise interference effectively because of the sign operation on the instantaneous gradient based on constant modulus algorithm (CMA) cost function. Meanwhile, the excess mean square error can be further reduced by use of a variable step size algorithm based on the iterative times and the reliability of the output signal. Simulation results show that, the variable step size blind equalization based on sign gradient algorithm has the fastest convergence rate and the lowest steady state residue error compared with the fractional low order CMA, the nonlinear transform CMA and stop-and-go CMA.