Título:	Identificação de sinais radar pulsados por meio de transformada de Wavelet contínua e redes neurais artificiais
Outros títulos:	Identification of pulsed radar signals by means of Wavelet continuous transform and artificial neural networks
Autor(es):	Molina, André Luiz Bandeira
Orientador(es):	Nascimento, Francisco Assis de Oliveira
Assunto:	Radar Wavelets (Matemática) Redes neurais (Computação)
Data de publicação:	19-Out-2010
Data de defesa:	15-Dez-2006
Referência:	MOLINA, André Luiz Bandeira. Identificação de sinais radar pulsados por meio de transformada de Wavelet contínua e redes neurais artificiais. 2006. 112 f., il. Dissertação (Mestrado em Engenharia Elétrica)-Universidade de Brasília, Brasília, 2006.
Abstract:	The deinterleaving of radar signal pulse sequences is an essential work for the radar identification in electronic warfare. Some ways of performing deinterleaving are based on temporal techniques, that use histogram and recursivity, and in techniques based on orthogonal transforms, that use spectral signatures for the detection of the sequences. Any of them require heavy computational power. However, the orthogonal transform techniques are more robust due to the properties of correlation minimization among the several signal fonts. The spectral signatures in the transformed space show a typical behavior more clearly. This work proposes a technique composed of three distinct phases. First, there is a preprocessing by which the whole interval of interest is segmented into sub intervals by means of downsampling and digital filtering. After that, it is used a time-frequency analysis by means of continuous wavelet transform, separating the spectral patterns of interest. Finally, it is implemented an identification process of spectral signature based on artificial neural networks. It is also described all the necessary mathematics concepts to the establishment of the contour conditions and elaboration of the dedicated procedures. The algorithms were developed in MATLAB for the simulation and validation of the models. The results of the computer simulation show a good performance and robustness of these algorithms when several sources of pulsed radar signals are present. The proposed model gives a good precision for the spectral signature, what makes the identification of radar signals possible by means of the artificial neural networks.
Unidade Acadêmica:	Faculdade de Tecnologia (FT) Departamento de Engenharia Elétrica (FT ENE)
Informações adicionais:	Dissertação (mestrado)—Universidade de Brasília, Faculdade de Tecnologia, Departamento de Engenharia Elétrica, 2006.
Programa de pós-graduação:	Programa de Pós-Graduação em Engenharia Elétrica
DOI:	http://dx.doi.org/10.26512/2006.12.D.5710
Aparece nas coleções:	Teses, dissertações e produtos pós-doutorado

Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.