A New Variable Step Size LMS Algorithm with Application to Active Noise Control

Y Wang,C Zhang,ZH Wang
DOI: https://doi.org/10.1109/icassp.2003.1200034
2004-01-01
Abstract:A new adaptive step size adjustment least mean square (LMS) algorithm is presented. The proposed algorithm modifies the existing LMS using the estimated output error as an important component for the modification of the step size. Experiment results demonstrate that application of the new algorithm leads to a significant gain in SNR (signal-to-noise ratio), thus visibly reduces the output error and greatly improves the convergence speed in the context of adaptive active noise cancellation.
What problem does this paper attempt to address?