A Utilização de Redes Neurais Artificiais na Estimação da Cobertura do Sinal de Televisão Digital
Carregando...
Arquivos
Data
2009-08-21
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Goiás
Resumo
THIS works presents information about signal intensity obtained on field and
from simulations for one-seg and full-seg receptions, the city of Goi ania. The
values obtained from measurements were used for a comparisson among propagation
models that are presented in literature, and the goal is to determine the real
condition of digital TV signal in the region of Goi ania.
The propagation models presented are available in literature and can be implemented
in digital transmission system. The studied models were Free Space model,
Log-Distance model, Hata model and ITU-R P.1546-1 method, and the objective
was to determine the signal intensity of digital television transmission in the city
of Goi ania (RAPPAPORT, 1996) (UNION, 2003). Focusing on the development of a
tool for signal intensity estimation, some researches were done about neural networks
theory and its applications. Perceptron and Multilayer Perceptron were the
analised architectures, emphasyzing on the last one and on its supervisioned trainning
through the backpropagation error algorithm (HAYKIN, 2001). The Brazilian
Digital Television System was described by reference rules made by Associa¸c ao
Brasileira de Normas T´ecnicas, which has detailed its transmission system and
reception devices (T´ECNICAS, 2008a) (T´ECNICAS, 2008h).
Measurements of signal intensity for one-seg and full-seg reception methods
were made on field in the region of Goi ania. These measurements followed the
sugestions presented by Report ITU-R BT. 2035-1 and it used a radiofrequency
analyzer and a Global Positioning System (GPS). With the obtained data, the
digital signal covering situation in the city of Goi ania was mapped, which revealed
a lower intensity level to the studied models. A tool for signal intensity estimation
was developed using Artificial Neural Networks, which was trained with the dada
obtained from the performed measurements. This tool was used to obtain the
signal intensity for several proposed scenarios. The signal intensity estimation for
the scenario that has tree density and target absence distinguished as the one that
was closest to the reality of Goi ania, which is a consequence of high density of
trees.
Descrição
Citação
SILVA, Douglas Dias da. The Use of Artificial Neural Networks in the Estimation of Coverage Digital TV Signal. 2009. 193 f. Dissertação (Mestrado em Engenharia) - Universidade Federal de Goiás, Goiânia, 2009.