Integrated CMOS Analog Neural Network Ability to Linearize the Distorted Characteristic of HPA Embedded in Satellites

Author

Gatet, Laurent ; Tap-Beteille, H. ; Roviras, Daniel ; Gizard, Francis

Author_Institution

CNES, Toulouse

fYear

2008

fDate

23-25 Jan. 2008

Firstpage

502

Lastpage

505

Abstract

New satellites generations have regenerative payload. Received signals are demodulated in order to be processed, before being modulated again and amplified on-board. For correcting non-linearities due to high power amplifier (HPA) operating near saturation, a predistortion system based on neural networks (NNs) was developed. For size, consumption and bandwidth purposes, the multi-layer perceptron (MLP) type NNs were implemented in a 0.6 mum CMOS ASIC. This paper presents the linearization ability of the analog integrated NN to correct the AM/AM distortion due to the HPA for changing simulation conditions (temperature drifts, ageing variations).

Keywords

CMOS analogue integrated circuits; application specific integrated circuits; multilayer perceptrons; power amplifiers; space vehicle electronics; AM-AM distortion; CMOS ASIC; ageing variations; distortion characteristics; high power amplifier; integrated CMOS analog neural network; linearization; multilayer perceptron; size 0.6 mum; telecommunication satellites; temperature drifts; Application specific integrated circuits; Bandwidth; High power amplifiers; Multilayer perceptrons; Neural networks; Payloads; Predistortion; Satellites; Signal processing; Temperature; CMOS Analog Integrated Circuits; Multi-Layer Perceptrons; Neural Network Architecture; Nonlinear Distortion; Predistorsion;

fLanguage

English

Publisher

ieee

Conference_Titel

Electronic Design, Test and Applications, 2008. DELTA 2008. 4th IEEE International Symposium on

Conference_Location

Hong Kong

Print_ISBN

978-0-7695-3110-6

Type

conf

DOI

10.1109/DELTA.2008.49

Filename

4459602