Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Achievable Rate Optimization for Large Stacked Intelligent Metasurfaces Based on Statistical CSI

Anastasios Papazafeiropoulos, Pandelis Kourtessis, Symeon Chatzinotas, Dimitra I. Kaklamani, Iakovos S. Venieris

TL;DR

The paper tackles maximizing downlink achievable rate in a multiuser SIM-aided MIMO system under statistical CSI to reduce overhead. It develops a tractable SE expression using the UaTF bound and formulates a non-convex optimization over SIM phase shifts and transmit power, solved by an alternating optimization framework. The SIM optimization uses projected gradient ascent with unit-modulus constraints, while power allocation relies on a weighted MMSE reformulation, both offering low computational complexity and convergence to local optima. Numerical results show that larger SIMs and more layers improve sum SE, and statistical CSI offers a practical overhead-time trade-off with performance close to instantaneous CSI in many scenarios.

Abstract

Stacked intelligent metasurface (SIM) is an emerging design that consists of multiple layers of metasurfaces. A SIM enables holographic multiple-input multiple-output (HMIMO) precoding in the wave domain, which results in the reduction of energy consumption and hardware cost. On the ground of multiuser beamforming, this letter focuses on the downlink achievable rate and its maximization. Contrary to previous works on multiuser SIM, we consider statistical channel state information (CSI) as opposed to instantaneous CSI to overcome challenges such as large overhead. Also, we examine the performance of large surfaces. We apply an alternating optimization (AO) algorithm regarding the phases of the SIM and the allocated transmit power. Simulations illustrate the performance of the considered large SIM-assisted design as well as the comparison between different CSI considerations.

Achievable Rate Optimization for Large Stacked Intelligent Metasurfaces Based on Statistical CSI

TL;DR

The paper tackles maximizing downlink achievable rate in a multiuser SIM-aided MIMO system under statistical CSI to reduce overhead. It develops a tractable SE expression using the UaTF bound and formulates a non-convex optimization over SIM phase shifts and transmit power, solved by an alternating optimization framework. The SIM optimization uses projected gradient ascent with unit-modulus constraints, while power allocation relies on a weighted MMSE reformulation, both offering low computational complexity and convergence to local optima. Numerical results show that larger SIMs and more layers improve sum SE, and statistical CSI offers a practical overhead-time trade-off with performance close to instantaneous CSI in many scenarios.

Abstract

Stacked intelligent metasurface (SIM) is an emerging design that consists of multiple layers of metasurfaces. A SIM enables holographic multiple-input multiple-output (HMIMO) precoding in the wave domain, which results in the reduction of energy consumption and hardware cost. On the ground of multiuser beamforming, this letter focuses on the downlink achievable rate and its maximization. Contrary to previous works on multiuser SIM, we consider statistical channel state information (CSI) as opposed to instantaneous CSI to overcome challenges such as large overhead. Also, we examine the performance of large surfaces. We apply an alternating optimization (AO) algorithm regarding the phases of the SIM and the allocated transmit power. Simulations illustrate the performance of the considered large SIM-assisted design as well as the comparison between different CSI considerations.
Paper Structure (10 sections, 2 theorems, 33 equations, 4 figures)

This paper contains 10 sections, 2 theorems, 33 equations, 4 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. Downlink Data Transmission
  4. Problem Formulation and Optimization
  5. Problem Formulation
  6. SIM Optimization
  7. Power Optimization
  8. Numerical Results
  9. Conclusion
  10. Proof of Proposition \ref{['propositionGradient']}

Key Result

Proposition 1

The achievable SINR of UE $k$ for a given SIM during the downlink transmission is provided by gamma2.

Figures (4)

  • Figure 1: A SIM-aided MIMO system.
  • Figure 2: Achievable sum SE of the large SIM-aided MIMO architecture with respect to the number of meta-atoms $N$.
  • Figure 3: Achievable sum SE of the large SIM-aided MIMO architecture with respect to the number of meta-surfaces $L$.
  • Figure 4: Achievable sum SE of the large SIM-aided MIMO architecture with respect to the number of iterations for $5$ different randomly generated initial points.

Theorems & Definitions (5)

  • Proposition 1
  • proof
  • Proposition 2
  • proof
  • Remark 1