Improved frequency identification via an adaptive filters cascade

A problem of frequency identification for a biased sinusoidal signal is considered. An identification algorithm is proposed which ensures exponential convergence for a case of an input signal with no noise and additionally allows to estimate an amplitude and a bias of the input signal. For a case of a noised input signal an adaptive cascade of band-pass filters is proposed coupled with the proposed identification algorithm. The proposed cascade can significantly improve identification performance, attenuate oscillations in an identification error signal and reduce an identification error bias. An efficiency of the proposed scheme is illustrated via numerical simulations.