Classificação Automática de Modulações utilizando Redes Neurais Artificiais com regularização Bayesiana e algoritmo de retropropagação de Levenberg-Marquardt

Myke D. M. Valadão, Waldir Silva, André L. A. Costa, Celso Barbosa Carvalho, Diego A. Amoedo, Lucas Cordeiro, Eddie B de Lima Filho, Antonio M. C. Pereira

DOI: 10.14209/SBRT.2020.1570649633

Evento: XXXVIII Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2020)

Keywords: Classificação automática de modulações Redes neurais artificiais Rádio cognitivo

Abstract

With the increase in demand for frequency spectrum, it was necessary to search for more efficient ways to allocate users in the spectrum. The cognitive radio senses the spectrum and dynamically allocates users in unused spaces. The techniques for automatic classification of modulation have come to provide information that assist in spectrum sensing. For this work, characteristics were extracted from signals modulated in passband in differents SNR. The classifier used was a perceptron network with Bayesian regularization and Levenberg-Marquardt backpropagation algorithm. For the proposed method, obtained accuracy that varies from 74.8(%) to 95.5(%).

Download