Abstract:

In broad-band communication systems, the practical implementation of digital predistortion suffers from the sampling-rate limitation of analog-to-digital converters. In order to solve this problem, an under-sampling digital predistortion method is proposed. First, the power amplifier’s orthogonal model parameters are extracted by offline identification with under-sampling rate. Second, the simplified predistorter with Nyquist rate is obtained according to a certain algorithm. Finally, an adaptive direct-learning algorithm is operated online. The proposed method has the characteristics of simple predistorter model structure and efficient under-sampling realization. Theoretical analysis and simulation results demonstrate that the method reduces the sampling-rate requirement of the analog-to-digital converter effectively and optimizes the linearization condition of the simplified predistorter model, thus it can greatly suppress the in-band distortion and spectrum regrowth caused by the nonlinearities of the power amplifier.