A nonlinear modified constant modulus algorithm for blind equalization

In this paper, a nonlinear modified constant modulus algorithm (NMCMA) is proposed by applying a nonlinear transfer function (NTF) into modified constant modulus algorithm (MCMA), where the NTF makes the decision region of its cost function sharper so that NMCMA is more robust against the convergency error. The single layer in NMCMA simultaneously guarantees a comparably speedy convergency. Computer simulations show that, NMCMA achieves a lower convergency MSE than MCMA by approximately 15dB for 16-QAM symbols, by approximately 10dB for 32-QAM symbols.