ADAPTIVE ALGORITHMS FOR JOINT TIME-DELAY ESTIMATION AND IIR FILTERING

In this paper, a new class of gradient-based algorithms for joint adap tive time delay estimation and IIR filtering for the recursive identif ication of systems with long pure delays is introduced, The algorithms update an adaptive delay estimate and the coefficients of an IIR filt er using familiar gradient-descent methods, In addition, interpolation of the input sequence is used to obtain delays that are not constrain ed to integer numbers of samples, Three algorithms using a direct form filter parametrization and one using a normalized lattice filter are derived, and computer simulations are used to demonstrate their conver gence and tracking properties.