Joint polynomial and look-up-table power amplifier linearization scheme

Abstract

Power amplification is one of the crucial functions in wireless communications systems. Unfortunately, power amplifiers often encounter the problem of severe nonlinear distortion. This paper studies two types of predistortion linearization schemes for power amplifiers: polynomial and look-up-table (LUT). With limited complexity in both linearization schemes, the polynomial scheme suffers curve-fitting distortion, while the LUT scheme suffers granular distortion, i.e., distortion at those parts of a power amplifier signal where its input amplitudes are between the values specified by the LUT entries. To reach adjacent channel power ratio (ACPR) of up to 60 dBc or even higher for signals at a power amplifier output usually requires a very complicated system with the involved algorithms converging rather slowly. However, the nonlinear distortions remaining in the polynomial scheme and in the LUT scheme are quite different. Therefore, the two linearization schemes are capable of canceling the remaining distortion for each other. With this in mind, we propose a joint polynomial and LUT power amplifier linearization scheme where the polynomial scheme effectively minimizes the granular distortion that the LUT scheme might suffer. Baseband simulations are conducted to justify the proposed joint linearization scheme.