Adaptive algorithm for IIR filtering based on matrix iterative techniques

The authors present an adaptive algorithm based on matrix iterative methods to identify the parameters of an infinite impulse response (IIR) system. The objective is to estimate the system parameters with fast convergence using as few data samples as possible. Nonadaptive matrix iterative algorithms have been proposed by M. Tummala (1990, 1991) which exploit the structure of the covariance matrix by splitting it into a Toeplitz and a residual part. The authors propose extensions to that work which lead to a data adaptive algorithm for IIR filters. Simulation results are presented to show the convergence performance of the algorithm