A Predistortion Algorithm Based on Accurately Solving the Reverse Function of Memory Polynomial Model

This letter presents a predistortion algorithm which identifies the power amplifier (PA) model and accurately calculates the predistortion function from the model by a new method. The method constructs a univariate polynomial and finds its roots by QR factorization. Although additional calculations are required, the new algorithm can compensate the nonlinearity of the PA more precisely and reduce the normalized mean square error (NMSE) performance of the new algorithm to the NMSE of the PA modeling. Simulation results demonstrate that the new algorithm outperforms the conventional algorithm by 4dB in the adjacent channel leakage ratio (ACLR) performance.