An Adaptive Noise Canceller Using Error Nonlinearities in the LMS Adaptation

This paper introduces a new adaptive noise canceller (ANC) using a modified least mean-square (LMS) algorithm that applies nonlinearities to the error signal in the LMS update equation. The proposed algorithm for ANCs can be viewed as a variable step-size LMS algorithm, in which the step-size is inversely proportional to the square norm of the error vector which has an increasing length. With an appropriate choice of the dimensionless adaptation constant step-size, a trade-off between speed of convergence and misadjustment can be achieved. The proposed algorithm is simulated using different noise power levels for both stationary and nonstationary noise environments. Simulation results, carried out using a real speech, clearly demonstrate the superiority of the proposed algorithm over many other algorithms in achieving small values of steady-state excess mean-square error with high rates of convergence in stationary as well as nonstationary noise environments.