An New Adaptive Digital Predistortion Technology Based on RBF-Neural Network

摘要

To overcome the inherent nonlinearity of high power amplifier, we have to make a linearization process. First, this paper makes a comparison between BP and RBF Neural Network. And then on the base of description of the basic principle about digital predistortion, the paper propose a adaptive digital predistortion technology based on RBF-Neural Network, which can improve amplitude-distortion and phasedistortion at the same time, having an obvious impromerit on ACPR and CLIPPING. Last, the paper use a 16QAM signal to make a matlab simulation, the result also prove this technology have a good performance.