New forms of LS lattice algorithms and an analysis of their round-off error characteristics

Abstract

New formulas are presented for direct updating of the reflection coefficients in the a priori and a posteriori forms of the least squares (LS) lattice algorithms. An analysis of the numerical characteristics of the new LS lattice forms is given and a comparison is made with the numerical characteristics of the conventional LS lattice algorithms. Computer simulation results are provided on the numerical accuracy of conventional and new LS lattice algorithms for several values of finite word length.