A Novel Undersampling Digital Predistortion for Power Amplifier Using Orthogonal Polynomial

With the advancement of modern communication, the utilized bandwidth has been progressively increasing, imposing higher demands on the ADC sampling rate in digital predistortion (DPD) systems. Consequently, undersampling DPD techniques have emerged as a solution to alleviate the hardware burden. Among these techniques, spectral extrapolation is commonly employed in undersampling DPD technology to restore the bandwidth-limited output signal using extrapolation algorithms. However, as the acquisition bandwidth decreases, the performance of spectral extrapolation deteriorates. By analyzing the deficiencies of traditional spectral extrapolation models from the perspective of numerical stability. In this paper, our objective was to mitigate this issue by incorporating orthogonal polynomials into the conventional basis functions. This modification successfully reduced the condition number of the inverse matrix in the extrapolation operator. The simulation results show that the proposed model exhibits excellent extrapolation performance compared to conventional models. When the sampling rate is adjusted to match the original bandwidth, the extrapolation effect remains superior, with the NMSE improving by nearly 20 dB.