Title :
Mixed analog/digital implementation of MLP NN for predistortion
Author :
Langlet, F. ; Roviras, D. ; Mallet, A. ; Castanié, F.
Author_Institution :
TeSA, ENSEEIHT, Toulouse, France
fDate :
6/24/1905 12:00:00 AM
Abstract :
In this paper, we present an adaptive predistortion technique for a regenerative satellite downlink with a 100 Mbits/s data rate. This link consists of an I/Q 16-QAM modulator followed by a solid state power amplifier (SSPA) on board the satellite. The memoryless amplifier gives two kinds of distortion: amplitude distortion (AM/AM conversion) and phase distortion (AM/PM conversion). Predistortion is realized with multi-layer perceptron (MLP) neural networks (NN). We present a mixed analog/digital implementation of the algorithm. We analyze the implementation imperfection effects for comparison with the theoretical algorithm
Keywords :
CMOS integrated circuits; adaptive signal processing; amplitude modulation; electric distortion; mixed analogue-digital integrated circuits; multilayer perceptrons; neural chips; phase modulation; satellite links; 100 Mbit/s; AM/AM conversion; AM/PM conversion; CMOS technology; I/Q 16-QAM modulator; MLP NN; SSPA; adaptive predistortion technique; algorithm imperfection effects; amplitude distortion; computer simulations; memoryless amplifier; mixed analog/digital implementation; multi-layer perceptron neural networks; phase distortion; regenerative satellite down link; solid state power amplifier; Algorithm design and analysis; Downlink; Multi-layer neural network; Multilayer perceptrons; Neural networks; Phase distortion; Power amplifiers; Predistortion; Satellites; Solid state circuits;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1007596
