FACTORIZED ALL-PASS IIR ADAPTIVE NOTCH FILTERS

J. E. COUSSEAU; P. D. DOÑATE; Y. LIU

Quebec

Congreso; International Conference on Acoustics, Speech and Signal Processing; 2004

IEEE Signal Processing Society

A new family of IIR adaptive notch filters is presented. Thefamily proposed is based on a second-order factorization ofthe all-pass transfer function that forms the multiple notchfilter. These realizations represent an extension of a previousad-hoc scheme for adaptive notch filtering, that avoida high-order polynomial root finding in order to obtain theunknown frequencies of interest. An interesting aspect relatedto these novel algorithms is the fact that they introducea different compromise between bias and SNR if comparedwith previous realizations available in the literature. Specifically,lower bias than in other approaches for low SNRcan be achieved using the new realizations. This property isparticularly attractive for multiple sinusoids estimation andtracking. In addition to the algorithm presentation, a discussionof the different properties and characteristics (stationarypoints, convergence) is also included. Also, computersimulations are presented in order to illustrate the expectedperformance of the adaptive filters proposed.