Abstract
A new approach to solving the adaptive ITR lattice filtering problem is presented here. In this approach, correlation ( or secondary ) data are used to estimate the parameters of the unknown filter. As a result, the filtering algorithm derived will treat equation error model and output error model in the exactly same way; and the criterion function of this algorithm is always quadratic. By using a minimum- phase recursive multichannel AR process lattice estimator, a minimum-phase adaptive IIR lattice filtering algorithm can be readily realized.