[Paper Review] Enabling Panoramic Full-Angle Reflection via Aerial Intelligent Reflecting Surface

Haiquan Lu, Yong Zeng+2 more|arXiv (Cornell University)|Jan 21, 2020

Advanced Wireless Communication Technologies13 references20 citations

TL;DR

This paper proposes a 3D aerial intelligent reflecting surface (AIRS) architecture to enable 360° panoramic full-angle signal reflection, overcoming the 180° half-space limitation of terrestrial IRS. By jointly optimizing transmit beamforming, AIRS placement, and phase shifts via closed-form solutions and analog beamforming-inspired suboptimal design, the system maximizes worst-case SNR, achieving significant performance gains over heuristic deployment schemes in urban environments.

ABSTRACT

This paper proposes a new three dimensional (3D) networking architecture enabled by aerial intelligent reflecting surface (AIRS) to achieve panoramic signal reflection from the sky. Compared to the conventional terrestrial IRS, AIRS not only enjoys higher deployment flexibility, but also is able to achieve 360$^\circ$ panoramic full-angle reflection and requires fewer reflections in general due to its higher likelihood of having line of sight (LoS) links with the ground nodes. We focus on the problem to maximize the worst-case signal-to-noise ratio (SNR) in a given coverage area by jointly optimizing the transmit beamforming, AIRS placement and phase shifts. The formulated problem is non-convex and the optimization variables are coupled with each other in an intricate manner. To tackle this problem, we first consider the special case of single-location SNR maximization to gain useful insights, for which the optimal solution is obtained in closed-form. Then for the general case of area coverage, an efficient suboptimal solution is proposed by exploiting the similarity between phase shifts optimization for IRS and analog beamforming for the conventional phase array. Numerical results show that the proposed design can achieve significant performance gain than heuristic AIRS deployment schemes.

Motivation & Objective

• Address the fundamental limitations of terrestrial intelligent reflecting surfaces (IRS), including restricted 180° coverage, deployment inflexibility, and high signal loss due to multiple reflections.
• Overcome these issues by introducing a 3D aerial IRS (AIRS) architecture mounted on UAVs or balloons to enable full-sky, 360° panoramic signal reflection with higher line-of-sight (LoS) probability.
• Maximize the worst-case signal-to-noise ratio (SNR) across a target area by jointly optimizing transmit beamforming, AIRS position, and phase shifts.
• Provide analytical insights through a closed-form solution for the single-location SNR maximization problem, revealing that optimal AIRS height depends on the ratio of height to source-destination distance.
• Develop an efficient suboptimal solution for area coverage by exploiting the analogy between IRS phase shift optimization and analog beamforming in phase arrays.

Proposed method

• Formulate a non-convex optimization problem to maximize the minimum SNR across a ground area, jointly optimizing transmit beamforming, AIRS location (3D coordinates), and passive phase shifts.
• Derive a closed-form optimal solution for the special case of single-location SNR maximization by analyzing the SNR expression and minimizing a derived function of the form $ f( heta) = ( heta^2 + ho^2)(( heta - 1)^2 + ho^2) $, where $ ho = H / orm{f{ ilde{w}}} $.
• Use cubic equation analysis (via discriminant $ riangle $) to solve the first-order derivative $ f'( heta) $, identifying critical points and determining the global minimum of the SNR function.
• For the general area coverage problem, propose a suboptimal algorithm inspired by analog beamforming in phased arrays, leveraging the similarity between IRS phase shift optimization and beamforming beamsteering principles.
• Apply coordinate transformation $ heta = ilde{ heta} + 1/2 $ and analyze three cases based on the discriminant $ riangle = (b/2)^2 + (a/3)^3 $, where $ a = ho^2 - 1/4 $, to determine optimal phase shift configurations.
• Validate the design through numerical results comparing the proposed method with heuristic AIRS deployment schemes, demonstrating superior SNR performance and robustness in complex urban environments.

Experimental results

Research questions

• RQ1 Can an aerial intelligent reflecting surface (AIRS) achieve 360° panoramic full-angle reflection, overcoming the 180° angular limitation of terrestrial IRS?
• RQ2 How does the optimal placement of AIRS depend on system parameters such as height and source-destination distance in a single-user scenario?
• RQ3 What is the closed-form solution for maximizing SNR at a single ground location via joint optimization of AIRS position and phase shifts?
• RQ4 How can the complex non-convex optimization problem for area-wide SNR maximization be efficiently solved when variables are tightly coupled?
• RQ5 To what extent does the proposed AIRS design reduce the number of signal reflections and associated path loss compared to terrestrial IRS in urban environments?

Key findings

• The optimal AIRS location for single-point SNR maximization is determined solely by the ratio of AIRS height $ H $ to the distance between source and destination, with the optimal phase shift configuration derived in closed-form.
• For the single-location case, the minimum SNR is achieved when $ heta = 1/2 $ if $ ho > 1/2 $, or at $ heta = 1/2 imes (1 imes rac{1}{2} imes rac{1}{2}) $ in the case of $ ho < 1/2 $, with the minimum value occurring at symmetric points $ heta = rac{1}{2} imes rac{1}{2} imes rac{1}{2} $.
• The proposed suboptimal design for area coverage achieves significant performance gains over heuristic AIRS deployment schemes, as demonstrated by numerical results showing improved worst-case SNR and reduced signal loss.
• The AIRS architecture reduces the number of required reflections in urban environments from multiple hops (as in terrestrial IRS) to a single reflection, significantly mitigating signal attenuation.
• The cubic equation analysis reveals that the SNR function is monotonic decreasing then increasing when $ ho > 1/2 $, and has a U-shaped profile with two symmetric minima when $ ho < 1/2 $, enabling precise optimization.
• The second-order derivative analysis confirms that the SNR function is convex around the optimal point, ensuring the derived solution is globally optimal for the single-point case.