We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
A Novel NN-Predistorter Learning Method for Nonlinear HPA.
- Authors
Cui, Hua; Zhao, Xiang-Mo
- Abstract
The deficiency of predistortion performance exists in indirect NN (Neural Network)-predistorter learning methods for nonlinear high power amplifiers (HPAs), and direct NN-predistorter learning methods possess great computational complexity. To circumvent these problems, in this paper we propose a novel NN-predistorter learning method with its structure developed by using some properties of nonlinear operators and its corresponding algorithm derived by using an approximation formula. The proposed method is based on the identification of NN post-distorter of the HPA, and then directly implements the efficient Levenberg-Marquardt back propagation algorithm. Thus, compared with the direct NN-predistorter learning method, our proposed method reduces the computational complexity and still keeps slightly better predistortion performance. Theoretical analysis and simulation results also show our proposed method outperforms the indirect NN-predistorter learning method in the term of about 5 dB adjacent channel power ratio improvement.
- Subjects
NONLINEAR operators; ALGORITHMS; MACHINE learning; APPROXIMATION theory; BACK propagation; COMPUTATIONAL complexity
- Publication
Wireless Personal Communications, 2012, Vol 63, Issue 2, p469
- ISSN
0929-6212
- Publication type
Article
- DOI
10.1007/s11277-010-0144-z