On extended RLS lattice adaptive filter variants: error-feedback, normalized and array-based algorithms

Abstract

This paper develops several lattice structures for RLS orthonormally-based input data structures, including error feedback, normalized and array-based forms. All recursions are theoretically equivalent, however they tend to differ in performance under finite precision effects. As a result, we verify that compared to the standard extended lattice equations, the new variants do not improve robustness to quantization, unlike what is normally expected for FlR models.