Title :
Results on a direct digital receiver operated with fast learning networks
Author :
Mireux, O. ; Brault, J.-J. ; Bosisio, R.G.
Author_Institution :
Centre de Recherche Avancee en Micro-ondes et Electronique Spatiale, Ecole Polytech. de Montreal, Que., Canada
Abstract :
In this paper, learning algorithms are used to decode QPSK modulated signals in a direct conversion microwave/millimetre wave receiver using an application specific six port module. Two different algorithms, K-Means and Online Bayesian Network, are considered for operation of the decoder to recover IQ data from modulated signals. Bit Error Rate (BER) vs. Noise level (Eb/No) results are presented including the case where the Local Oscillator (LO) is not locked to carrier signal.
Keywords :
adaptive decoding; belief networks; error statistics; iterative decoding; learning (artificial intelligence); microwave receivers; millimetre wave receivers; multiport networks; quadrature phase shift keying; telecommunication computing; IQ data recovery; K-Means algorithm; QPSK modulated signals; QPSK signal decoding; application specific six port module; bit error rate; direct conversion microwave/millimetre wave receiver; direct digital receiver; fast learning networks; iterative method; learning algorithms; local oscillator; noise level; online Bayesian network algorithm; six-port direct receiver; Bayesian methods; Bit error rate; Calibration; Clustering algorithms; Decoding; Equations; Frequency; Local oscillators; Quadrature phase shift keying; Signal to noise ratio;
Conference_Titel :
Microwave Symposium Digest, 2002 IEEE MTT-S International
Conference_Location :
Seattle, WA, USA
Print_ISBN :
0-7803-7239-5
DOI :
10.1109/MWSYM.2002.1011665
