An approach to transform domain variable step-size LMS adaptive filter

This paper introduces a new transform domain variable step-size LMS algorithm, in order to deal with highly correlated inputs. In the algorithm the step-size rule is based on the weighting coefficients bias to variance trade-off. Performance of this method is promising, especially in nonstationary environment with abrupt change of unknown system. Computer simulation results are provided to support the proposed implementation of the new TDVSS LMS.