Research on Digital Predistortion of Power Amplifier Based on Innovation Adaptive Kalman Filter

Abstract

In this paper, the adaptive digital predistorter (DPD) based on full Volterra (FV) series is studied for nonlinearity compensation of power amplifier (PA). In order to overcome the weakness of the traditional Kalman filter (TKF), we propose an adaptive Kalman filter (AKF) algorithm based on innovation which does not require prior knowledge of measurement noise. The innovation is introduced into the Kalman filter to correct the Kaman gain by calculating the maximum likelihood optimal estimating of the innovation variance. With this method, the proposed scheme can effectively suppress the measurement noise and reduce the adverse effect of noises on the accuracy of coefficients identification. The simulation results show that the proposed AKF algorithm can effectively track the change of actual noise, and the accuracy of coefficients identification in noise environment is improved by about 20%-50% than that of TKF. Furthermore, the proposed AKF has a better performance in terms of power spectral density and bit error rate (BER) than TKF, which demonstrates the prospective and validity in nonlinearity compensation for PA.