A straightforward, general-purpose, robust adaptive IIR filter

Abstract

An adaptive IIR (Infinite Impulse Response) filter structure is presented which models an unknown linear system in the presence of significant colored additive noise far better than the methods proposed in the literature, as surveyed by the author. With the proposed structure the poles and zeroes of the transfer function of an unknown linear system can be readily identified. The identification of pole and zero locations provides profound insight into the nature of the environment (valuable for the improvement of active sensors). In computer simulations, the adaptive IIR filter structure presented in this paper was able to accurately model an unknown 6th-order system in an environment of 0 dB SNR (Signal-to-Noise Ratio) to a Signal <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> -to-Error <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> (S/E) ratio averaging 29 dB with colored additive noise. The S/E ratio achieved by the presented algorithm represents an improvement of over 2 orders of magnitude, compared to the methods proposed in the literature, as surveyed by the author. Additionally, the algorithm includes a method for determining the order of the system being identified.