Configuring Intelligent Reflecting Surface with Performance Guarantees: Blind Beamforming

TL;DR

This paper introduces a blind beamforming method for IRS that enhances SNR without channel estimation, using a statistical approach with random sampling, achieving quadratic SNR boost and demonstrated superior performance in field tests.

Contribution

Proposes a novel blind beamforming strategy for IRS based on conditional sample mean, avoiding channel estimation and achieving quadratic SNR improvement with polynomial samples.

Findings

Quadratic SNR boost with polynomial samples

Outperforms benchmark algorithms in field tests

Interprets blind beamforming as a least squares problem

Abstract

This work gives a blind beamforming strategy for intelligent reflecting surface (IRS), aiming to boost the received signal-to-noise ratio (SNR) by coordinating phase shifts across reflective elements in the absence of channel information. While the existing methods of IRS beamforming typically first estimate channels and then optimize phase shifts, we propose a conditional sample mean based statistical approach that explores the wireless environment via random sampling without performing any channel estimation. Remarkably, the new method just requires a polynomial number of random samples to yield an SNR boost that is quadratic in the number of reflective elements, whereas the standard random-max sampling algorithm can only achieve a linear boost under the same condition. Moreover, we gain additional insight into blind beamforming by interpreting it as a least squares problem. Field tests demonstrate the significant advantages of the proposed blind beamforming algorithm over the benchmark algorithms in enhancing wireless transmission.