الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 109 |  |  |
| | | from | 109 | | |
Abstract
Modern wireless communication networks are becoming increasingly
complex and diverse in their service requirements, which challenges the
current state-of-the-art wireless design methodology. Many engineering
fields, such as communications, speech and image processing, computer
vision, and robotics, have used deep learning since the last decades of the
20th century. It has proven particularly effective and useful in situations
where a rigorous mathematical model of the problem is difficult to elaborate.
Systems that are based on deep learning are more adaptable to distortions
than those that are not, and they approach the transmission of information in
a more holistic way than conventional systems. Considering wireless
communication systems, recent years have seen a great deal of attention paid
to machine learning applications in the upper layers, such as the deployment
of cognitive radio and self-organized networks or the management of
resources, while its application in the physical layer has been largely
ignored.
This thesis looks at the potential application of neural networks for
communication system physical layer optimization. The analysis is
particularly concerned with channel estimation for an orthogonal frequency
division multiplexing (OFDM) system. To be more specific, deep learning
gated recurrent unit (GRU) neural networks are used to present a new
framework for channel estimation. Initially, it is trained offline using
generated data sets, and thereafter it is used online to track the channel parameters, after which the data transmitted can be recovered. For the
purpose of determining the performance of the proposed estimator, three
alternative deep learning optimization techniques are used to test it, namely,
stochastic gradient descent with momentum (SGDm), root mean square
propagation (RMSProp), and adaptive moment (Adam). It is also compared
to other commonly used estimators, such as least squares (LS) and minimum
mean square error (MMSE). In addition, the proposed estimator is compared
with two existing models. Deep learning GRU neural network-based channel
state estimator, which are capable of learning and generalizing rapidly, are
shown to outperform the comparable estimators when just a few pilots are
available. In addition, there is no need for prior knowledge of channel
statistics. So, estimating OFDM communication system channel states using
the proposed estimator appears promising.