A New Variable Step-Size LMS Algorithm and Its Applications

A new variable step-size LMS adaptive filtering algorithm is presented in this paper,which adjusts its step-size according to the cross-correlation function of relative output error signal.Computer simulation results confirm the theoretical analysis and show that this algorithm have faster convergence rate,smaller misadjustment,and better stability.Also,it can achieve better performance in lower SNR situation.The superiority of this algorithm is that it has smaller computational complexity and is not influenced by existing uncorrelated noise.Thus it can be applied perfectly in practical signal processing.