Deep learning for energy-efficient wireless communications and spectrum management.
Access rightsNo access - contact email@example.com
Xing, Yuan, 1992-
MetadataShow full item record
In the past couple of decades, wireless communication has undergone rapid development. The current fourth generation and upcoming fifth generation wireless technologies promise us an ultra-fast data rate. However, a lot of energy is sacrificed in order to guarantee high quality communication. Therefore, energy-efficient wireless communication has been widely explored under the background of scarce energy resource and environmental-friendly data transmission. In order to assure the fast communication speed and network reliability, the structures of wireless communication systems become more and more complicated. The rational resource allocation in the sophisticated communication systems is a tough problem that urgently needs to be solved. The conventional communication theories exhibit limitations in fulfilling the perfect resource allocation in the systems. Nevertheless, Deep Learning methods are expert at solving sophisticated optimization problems. The key advantages of Deep Learning are the efficient learning of an enormous amount of data and the precise analysis for the hidden distribution. Therefore, Deep Learning methods can be used to solve complicated but useful energy efficiency optimization problems in wireless communication systems. This dissertation first explores an energy-efficient wireless communication system: the Simultaneous Wireless Information and Power Transfer system. Specifically, the wireless transmitters stably communicate with the information receivers, while there are several energy harvesters. The harvesters can take the electromagnetic waves as the energy source in order to charge the low power Internet of things. Deep Learning algorithms are utilized to optimize the wireless information and power transfer strategies. Second, this dissertation discusses another energy-efficient wireless communication system: a multiuser downlink Orthogonal Frequency Division Multiple Access data transmission system. In the system, the base station aims to achieve the highest communication quality with the least energy consumption. Deep Learning algorithms are applied to accomplish the energy-efficient wireless transmission. In summary, this dissertation investigates the usefulness of Deep Learning algorithms to boost the performance of two energy-efficient wireless communication systems, the Simultaneous Wireless Information and Power Transfer system and multiuser downlink Orthogonal Frequency Division Multiple Access data transmission system. The numerical results prove the excellence of Deep Learning methods in solving the optimization problems in energy-efficient wireless communication systems.