Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

WMMSE-Based Rate Maximization for RIS-Assisted MU-MIMO Systems

Hyuckjin Choi, A. Lee Swindlehurst, Junil Choi

TL;DR

This paper proposes rate maximization techniques for both single-user and multiuser MIMO systems, based on the well-known weighted minimum mean square error (WMMSE) criterion, and reveals that the proposed rate maximization technique exhibits superior performance when compared to other benchmarks.

Abstract

Reconfigurable intelligent surface (RIS) technology, given its ability to favorably modify wireless communication environments, will play a pivotal role in the evolution of future communication systems. This paper proposes rate maximization techniques for both single-user and multiuser MIMO systems, based on the well-known weighted minimum mean square error (WMMSE) criterion. Using a suitable weight matrix, the WMMSE algorithm tackles an equivalent weighted mean square error (WMSE) minimization problem to achieve the sum-rate maximization. By considering a more practical RIS system model that employs a tensor-based representation enforced by the electromagnetic behavior exhibited by the RIS panel, we detail both the sum-rate maximizing and WMSE minimizing strategies for RIS phase shift optimization by deriving the closed-form gradient of the WMSE and the sum-rate with respect to the RIS phase shift vector. Our simulations reveal that the proposed rate maximization technique, rooted in the WMMSE algorithm, exhibits superior performance when compared to other benchmarks.

WMMSE-Based Rate Maximization for RIS-Assisted MU-MIMO Systems

TL;DR

This paper proposes rate maximization techniques for both single-user and multiuser MIMO systems, based on the well-known weighted minimum mean square error (WMMSE) criterion, and reveals that the proposed rate maximization technique exhibits superior performance when compared to other benchmarks.

Abstract

Reconfigurable intelligent surface (RIS) technology, given its ability to favorably modify wireless communication environments, will play a pivotal role in the evolution of future communication systems. This paper proposes rate maximization techniques for both single-user and multiuser MIMO systems, based on the well-known weighted minimum mean square error (WMMSE) criterion. Using a suitable weight matrix, the WMMSE algorithm tackles an equivalent weighted mean square error (WMSE) minimization problem to achieve the sum-rate maximization. By considering a more practical RIS system model that employs a tensor-based representation enforced by the electromagnetic behavior exhibited by the RIS panel, we detail both the sum-rate maximizing and WMSE minimizing strategies for RIS phase shift optimization by deriving the closed-form gradient of the WMSE and the sum-rate with respect to the RIS phase shift vector. Our simulations reveal that the proposed rate maximization technique, rooted in the WMMSE algorithm, exhibits superior performance when compared to other benchmarks.
Paper Structure (16 sections, 77 equations, 12 figures, 2 tables, 2 algorithms)

This paper contains 16 sections, 77 equations, 12 figures, 2 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. System and Channel Models
  4. MISO/SIMO Systems
  5. Problem Objective
  6. WMMSE-Based RIS phase shift optimization
  7. Sum-Rate Maximization
  8. Gradient of the Achievable Rate
  9. Evaluation of First Gradient
  10. WMSE Minimization
  11. CNN-based Approach
  12. Adaptation to SU-MIMO Systems
  13. Simulation Results
  14. Experiments for SU-MIMO Systems
  15. Experiments for MU-MIMO Systems
  16. ...and 1 more sections

Figures (12)

  • Figure 1: System model considered in this paper. The BS has $M$ antennas, the RIS has $L$ elements, and $K$ UEs have $N$ antennas each.
  • Figure 2: Structure of CNN-WMMSE network.
  • Figure 3: Flowchart of CNN-WMMSE algorithm.
  • Figure 4: Simulation scenario for the SU-MIMO and MU-MIMO systems. The UEs are uniformly distributed in a 50m $\times$ 30m area, where the height of BS is 35 m, and the height of RIS is 15 m.
  • Figure 5: SU-MIMO achievable rates versus number of RIS elements, with 8 BS antennas and 4 UEs.
  • ...and 7 more figures

Theorems & Definitions (2)

  • proof
  • proof