A SAW neural network processor for classification of MSK and BPSK modulations

Author

Kalinin, V. ; Kavalov, D.

Volume

1

fYear

2002

fDate

8-11 Oct. 2002

Firstpage

263

Abstract

A SAW neural network processor for automatic recognition of two types of digital passband modulations is proposed in this paper. A feed-forward network with six hidden neurons is trained to recognize BPSK and MSK signals and the corresponding SAW NN modulation classifier is synthesised. Its performance is tested in the presence of additive white Gaussian noise. The influence of second-order effects in the SAW filters on the performance of the processor is also investigated. The results of this simulation show 95-99% probability for correct recognition at signal-to-noise ratio 10-12 dB.

Keywords

AWGN; digital signal processing chips; feedforward neural nets; minimum shift keying; neural chips; neural net architecture; phase shift keying; signal classification; surface acoustic wave filters; surface acoustic wave signal processing; BPSK; MSK; SAW filters; SAW neural network processor; additive white Gaussian noise; automatic recognition; classification; digital passband modulations; feed-forward network; hidden neurons; second-order effects; Binary phase shift keying; Digital modulation; Feedforward systems; Network synthesis; Neural networks; Neurons; Passband; Signal synthesis; Surface acoustic waves; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Ultrasonics Symposium, 2002. Proceedings. 2002 IEEE

ISSN

1051-0117

Print_ISBN

0-7803-7582-3

Type

conf

DOI

10.1109/ULTSYM.2002.1193398

Filename

1193398