Time Domain Delay Items Design for Memory Orthogonal Polynomial Predistorter

Abstract

The memory polynomial is commonly used in power amplifier (PA) modeling and predistorter (PD) design. However, the conventional memory polynomial and even memory orthogonal polynomial exhibit condition number increasing (eigenvalue spread) in covariance matrix when time domain delay items are included. In this paper, a method of designing time delay items for memory orthogonal polynomial is introduced. The method alleviates the eigenvalue spread problem effectively. Simulation results show that the new polynomial predistorter achieves much better predistortion performance based on indirect learning architecture (ILA) and normalized least mean square (NLMS) algorithm compared with the conventional memory orthogonal polynomial predistorter.