Performance comparison of LMS algorithms for acoustic noise cancellation

An acoustic noise cancellation (ANC) is a technique for removing additive noise from corrupted speech signal. This paper introduces a new Error Normalized Least Mean Square (ENLMS) algorithm in which the step size varies inversely with the squared norm of the error vector. The signed version of Normalized LMS (NLMS) algorithm requires a priori knowledge of a bound for the error magnitude, which is unknown in most applications. A very simple algorithm Signed Error Normalized LMS (SENLMS) is proposed which uses both these features- normalization by error function and a signed error. The proposed algorithm gives lower misadjustment and improved convergence speed, Mean Square Error (MSE), Signal-to-Noise Ratio (SNR), compared to LMS, NLMS and ENLMS algorithm.