Effects of demodulator errors on convergence time of look-up table based iterative algorithms for adaptive digital predistortion linearizer

Abstract

This work presents, for the first time, a simulation study of the effects of demodulator errors on look-up table (LUT) based iterative algorithms for adaptive digital predistortion (ADP) linearizer. We compare the iterative algorithms found in one-dimensional LUT, namely the polar, secant and linear methods, on the convergence time with and without demodulator errors in the feedback path of the ADP. Simulation shows that the secant method, although it has the fastest convergence, is sensitive to changes in demodulator errors, which could be corrected by adding an adaptation factor to its formula. The polar method is the least sensitive out of the three to demodulator errors, followed by the linear method.