Discrete Phase Reconfiguration of Intelligent Reflecting Surfaces Via Deep Learning

Abstract

Intelligent Reflective Surface (IRS), is a cost-effective technology suitable for achieving high spectrum and energy efficiency in future wireless communication systems. If the elements of an IRS are adjusted correctly, a power gain is provided that is directly proportional to the square of the number of elements. Configuring the elements of an IRS is a challenging task, and prior works have mostly assumed that the elements' phase shifts take continuous values. This paper investigates a wireless communication system in which the IRS phase shifts can only take a finite number of values. This study presents a Deep Learning approach for configuring discrete phase shifts of IRS elements. The proposed method uses the pilot signals reflected by the IRS as the model input and provides the optimum phase shift values as its output.