Traditional recursive least squares (RLS) adaptive filtering is widely used to estimate the impulse responses (IR) of an unknown system. Nevertheless, RLS estimator shows poor performance when tracking rapidly time-varying systems. In this paper, we propose a multi-layered (m-RLS) address concern. The m-RLS composed multiple estimators, each which employed and eliminate misadjustment previous l...

Total least squares parameter estimation is an alternative to least squares estimation though much less used in practice, partly due to the absence of eecient recursive algorithms or parallel architectures. Here it is shown how previously developed systolic algorithms/architectures for recursive least squares estimation can be used for recursive total least squares problems. Unconstrained as we...