Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learning-Based Joint Beamforming and Antenna Movement Design for Movable Antenna Systems

Caihao Weng, Yuanbin Chen, Lipeng Zhu, Ying Wang

TL;DR

The paper tackles sum-rate maximization for movable-antenna MA-enabled multi-receiver MIMO systems under imperfect CSI by jointly optimizing transmit beamforming and dual-sided antenna movement. It introduces a heterogeneous MADDPG framework with a dedicated beamforming agent and multiple MA movement agents, trained offline to deliver real-time policies that adapt to evolving CSI. A closed-form upper bound on per-user rate under imperfect CSI guides a reformulated problem (P2), enabling effective learning and coordination between agents. Simulation results show substantial sum-rate gains over benchmarks, with performance scaling with SNR and region size, and highlight the impact of channel estimation errors on achievable gains, underscoring the method’s robustness and practical relevance.

Abstract

In this paper, we investigate a multi-receiver communication system enabled by movable antennas (MAs). Specifically, the transmit beamforming and the double-side antenna movement at the transceiver are jointly designed to maximize the sum-rate of all receivers under imperfect channel state information (CSI). Since the formulated problem is non-convex with highly coupled variables, conventional optimization methods cannot solve it efficiently. To address these challenges, an effective learning-based algorithm is proposed, namely heterogeneous multi-agent deep deterministic policy gradient (MADDPG), which incorporates two agents to learn policies for beamforming and movement of MAs, respectively. Based on the offline learning under numerous imperfect CSI, the proposed heterogeneous MADDPG can output the solutions for transmit beamforming and antenna movement in real time. Simulation results validate the effectiveness of the proposed algorithm, and the MA can significantly improve the sum-rate performance of multiple receivers compared to other benchmark schemes.

Learning-Based Joint Beamforming and Antenna Movement Design for Movable Antenna Systems

TL;DR

The paper tackles sum-rate maximization for movable-antenna MA-enabled multi-receiver MIMO systems under imperfect CSI by jointly optimizing transmit beamforming and dual-sided antenna movement. It introduces a heterogeneous MADDPG framework with a dedicated beamforming agent and multiple MA movement agents, trained offline to deliver real-time policies that adapt to evolving CSI. A closed-form upper bound on per-user rate under imperfect CSI guides a reformulated problem (P2), enabling effective learning and coordination between agents. Simulation results show substantial sum-rate gains over benchmarks, with performance scaling with SNR and region size, and highlight the impact of channel estimation errors on achievable gains, underscoring the method’s robustness and practical relevance.

Abstract

In this paper, we investigate a multi-receiver communication system enabled by movable antennas (MAs). Specifically, the transmit beamforming and the double-side antenna movement at the transceiver are jointly designed to maximize the sum-rate of all receivers under imperfect channel state information (CSI). Since the formulated problem is non-convex with highly coupled variables, conventional optimization methods cannot solve it efficiently. To address these challenges, an effective learning-based algorithm is proposed, namely heterogeneous multi-agent deep deterministic policy gradient (MADDPG), which incorporates two agents to learn policies for beamforming and movement of MAs, respectively. Based on the offline learning under numerous imperfect CSI, the proposed heterogeneous MADDPG can output the solutions for transmit beamforming and antenna movement in real time. Simulation results validate the effectiveness of the proposed algorithm, and the MA can significantly improve the sum-rate performance of multiple receivers compared to other benchmark schemes.
Paper Structure (16 sections, 17 equations, 5 figures, 1 algorithm)

This paper contains 16 sections, 17 equations, 5 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. SYSTEM MODEL
  3. Scenario
  4. Channel Model and Communication Protocol
  5. Problem Formulation
  6. Heterogeneous MADDPG-Based Solution
  7. MA Agent
  8. State $s_{t,g}$
  9. Action $a_{t,g}$
  10. Reward $r_{t,g}$
  11. Beamforming Agent
  12. State $s_{t,b}$
  13. Action $a_{t,b}$
  14. Reward $r_{t,b}$
  15. Simulation Results
  16. ...and 1 more sections

Figures (5)

  • Figure 1: MA-enabled multi-receiver communication system.
  • Figure 2: The framework of proposed heterogeneous MADDPG.
  • Figure 3: Reward and sum-rate performance versus training time slots.
  • Figure 4: Sum-rate performance versus normalized region size.
  • Figure 5: Sum-rate performance versus SNR.

Theorems & Definitions (1)

  • proof