Digital Predistortion for Power Amplifier Based on Sparse Bayesian Learning

In this brief, a sparse-Bayesian-learning algorithm is applied to estimate the coefficients of the power amplifier (PA) behavioral models and inverse models from the view of probability. With this sparse learning method, the needed number of samplings can be reduced significantly. In addition, it also provides researchers with ideas that obtain the needed subspace of the preselected model. The performance of the algorithm is validated experimentally on a gallium nitride (GaN) PA, and the signal used to test the proposed approach is an Long Term Evolution (LTE) signal. A comparison with the state-of-the-art estimation algorithm in an open-loop digital-predistortion system is also presented, and the vast majority of tests show that the number of model coefficients is reduced by at least 50%.