Título: Equalizadores Neurais com Treinamento Baseado em Filtros de Kalman
Autores: Fernandes, Gabriel R.; Lima, Antonio C. de C.
Resumo: This work presents performance and convergence comparisons among different kinds of neural equalizers and standard decision feedback equalizers. The channels investigated here are considered linear, nonlinear, fixed and time-varying added with white Gaussian noise. The digital communication system employed in all simulations is BPSK and the learning algorithms applied for training are Kalman and RTRL.
Código DOI: 10.21528/CBRN2003-007
Artigo em PDF: 6CBRN_007.PDF
Arquivo BibTex: 6CBRN_007.bib