Paper Info
Title
Deep Learning for Large Intelligent Surfaces in Millimeter Wave and Massive MIMO Systems
Abstract
Large intelligent surfaces (LISs) are recently attracting increasing interest thanks to their promising coverage and data rate gains in future wireless systems. These surfaces comprise a massive number of nearly-passive elements that interact with the incident signals, for example by reflecting then, in a smart way that improves the wireless system performance. This smart interaction, however, requires acquiring large-dimensional channels between the LIS and the communicating transmitter/receiver. This channel estimation is associated with huge training overhead because of the massive number of LIS elements and the use of nearly-passive components, which represents a critical challenge for the LIS system operation. In this paper, a novel LIS architecture is proposed, where all the LIS elements are passive except for a few elements that are active (connected to the baseband of the LIS controller). Then, we develop a deep learning based solution where the LIS learns how to optimally interact with the incident signal given only the channels at the active elements. These channels represent the current state of the environment and the transmitter/receiver locations. The simulation results show that the developed solution can approach the upper bound that assumes perfect channel knowledge while requiring only a small fraction of the LIS elements to be active. This yields a promising solution for LIS systems from both energy efficiency and training overhead perspectives.
Year
DOI
Venue
2019
10.1109/GLOBECOM38437.2019.9013256
IEEE Global Communications Conference
Keywords
DocType
ISSN
large intelligent surface,smart reflect-array,beamforming,millimeter wave,deep learning
Conference
2334-0983
Citations 
PageRank 
References 
4
0.45
12
Authors
3
Name
Order
Citations
PageRank
Abdelrahman Taha1261.37
Muhammad Alrabeiah2323.88
Ahmed Alkhateeb3170867.18