Blind Beamforming for IRS Without Channel Estimation
Kaiming Shen, Zhi-Quan Tom Luo · 2023
This chapter comprises four main parts. First, we illustrate the necessity of configuring IRS blindly without any channel knowledge in real-world wireless networks. Second, we introduce a common heuristic method called RMS and further examine its fundamental aspect, i.e., how fast the SNR boost grows with the IRS size as well as the number of samples. Third, we propose a statistical method called CSM that guarantees a quadratic SNR boost. The main idea behind CSM is to explore the wireless environment by extracting its statistical features directly from the random samples of received signal power, thus not requiring any channel estimation whatsoever. Fourth, we connect CSM to the closest point projection and also to phase retrieval.