核心概念
The proposed adaptive blind beamforming algorithm enables effective configuration of intelligent surface (IS) without channel knowledge, especially in non-line-of-sight (NLoS) scenarios where existing blind beamforming methods may fail.
要約
The paper presents an adaptive blind beamforming algorithm for intelligent surface (IS) that can effectively configure the IS without requiring channel state information (CSI), particularly in non-line-of-sight (NLoS) scenarios where existing blind beamforming methods may fail.
The key insights are:
The existing blind beamforming algorithms, such as RFocus and CSM, rely on the assumption of a strong direct path between the transmitter and receiver. However, this assumption may not hold in NLoS scenarios, causing these algorithms to fail.
To address this issue, the proposed Grouped Conditional Sample Mean (GCSM) algorithm divides the reflective elements (REs) of the IS into three groups. This grouping strategy enables the algorithm to extract statistical features of the wireless environment and coordinate the phase shifts of the IS without explicit channel acquisition.
The GCSM algorithm is proven to work effectively for fading channels, unlike the existing blind beamforming methods that are limited to static channel assumptions.
The proposed algorithm is extended to the multi-user scenario, demonstrating its versatility and applicability in real-world wireless networks.
The key steps of the GCSM algorithm are:
Randomly divide the N REs of the IS into three groups.
For each group, obtain a dataset of random phase shift samples.
Extract the statistical features (empirical conditional averages) of the received signal power from the random samples.
Coordinate the phase shifts of the REs in each group to maximize the extracted statistical features.
Repeat steps 3-4 for the three groups in an alternating manner.
The GCSM algorithm is shown to outperform the existing blind beamforming methods, especially in NLoS scenarios, and requires fewer random samples to achieve good performance.
統計
The received signal power can be expressed as:
E[|Y|^2 | θn = φ] = 2P√(γ00γ0nγn0δ00δ0nδn0)/((1+δ00)(1+δ0n)(1+δn0))cos(φ-Δn) + σ^2 + P(γ00 + Σ_n γ0nγn0(δ0n+δn0+1)/((1+δ0n)(1+δn0)))
引用
"The RFocus and the CSM may fail to work in the non-line-of-sight (NLoS) channel case."
"To resolve this issue, this work advocates a novel adaptive blind beamforming algorithm, which divides and groups the REs to form a virtual direct path and thereby enables the empirical conditional average-based approach."