Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ADMM for Downlink Beamforming in Cell-Free Massive MIMO Systems

Mehdi Zafari, Divyanshu Pandey, Rahman Doost-Mohammady, César A. Uribe

TL;DR

The work tackles the high overhead of centralized downlink beamforming in cell-free MIMO by introducing an ADMM-based distributed optimization where each AP uses local CSI and shares a real interference vector with the central unit. The method decouples the SINR constraints via an interference consensus mechanism, enabling per-AP convex optimization and network-wide convergence. Results show the distributed approach converges within tens of iterations and delivers near-central performance while dramatically reducing fronthaul data, making it scalable for large networks. This approach offers a practical, low-latency beamforming solution for cell-free deployments with per-user QoS guarantees.

Abstract

In cell-free massive MIMO systems with multiple distributed access points (APs) serving multiple users over the same time-frequency resources, downlink beamforming is done through spatial precoding. Precoding vectors can be optimally designed to use the minimum downlink transmit power while satisfying a quality-of-service requirement for each user. However, existing centralized solutions to beamforming optimization pose challenges such as high communication overhead and processing delay. On the other hand, distributed approaches either require data exchange over the network that scales with the number of antennas or solve the problem for cellular systems where every user is served by only one AP. In this paper, we formulate a multi-user beamforming optimization problem to minimize the total transmit power subject to per-user SINR requirements and propose a distributed optimization algorithm based on the alternating direction method of multipliers (ADMM) to solve it. In our method, every AP solves an iterative optimization problem using its local channel state information. APs only need to share a real-valued vector of interference terms with the size of the number of users. Through simulation results, we demonstrate that our proposed algorithm solves the optimization problem within tens of ADMM iterations and can effectively satisfy per-user SINR constraints.

ADMM for Downlink Beamforming in Cell-Free Massive MIMO Systems

TL;DR

The work tackles the high overhead of centralized downlink beamforming in cell-free MIMO by introducing an ADMM-based distributed optimization where each AP uses local CSI and shares a real interference vector with the central unit. The method decouples the SINR constraints via an interference consensus mechanism, enabling per-AP convex optimization and network-wide convergence. Results show the distributed approach converges within tens of iterations and delivers near-central performance while dramatically reducing fronthaul data, making it scalable for large networks. This approach offers a practical, low-latency beamforming solution for cell-free deployments with per-user QoS guarantees.

Abstract

In cell-free massive MIMO systems with multiple distributed access points (APs) serving multiple users over the same time-frequency resources, downlink beamforming is done through spatial precoding. Precoding vectors can be optimally designed to use the minimum downlink transmit power while satisfying a quality-of-service requirement for each user. However, existing centralized solutions to beamforming optimization pose challenges such as high communication overhead and processing delay. On the other hand, distributed approaches either require data exchange over the network that scales with the number of antennas or solve the problem for cellular systems where every user is served by only one AP. In this paper, we formulate a multi-user beamforming optimization problem to minimize the total transmit power subject to per-user SINR requirements and propose a distributed optimization algorithm based on the alternating direction method of multipliers (ADMM) to solve it. In our method, every AP solves an iterative optimization problem using its local channel state information. APs only need to share a real-valued vector of interference terms with the size of the number of users. Through simulation results, we demonstrate that our proposed algorithm solves the optimization problem within tens of ADMM iterations and can effectively satisfy per-user SINR constraints.
Paper Structure (5 sections, 14 equations, 3 figures, 1 table, 1 algorithm)

This paper contains 5 sections, 14 equations, 3 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. Distributed Solution for Precoder Optimization
  4. Simulation Results
  5. Conclusion

Figures (3)

  • Figure 1: Illustration of the reference network architecture, with distributed APs connected to a central server and jointly serving many users.
  • Figure 2: SINR CDF for centralized vs distributed solution (Algorithm \ref{['alg:alg_admm']}) under the same SINR constraint $\gamma_k$ for all of the users across channel realizations. Left: $\gamma_k=15$ dB, Right: $\gamma_k=25$ dB.
  • Figure 3: Total transmit SNR for distributed approach (with 2 and 4 APs), and centralized solution (with 4 APs) under same SINR constraint $\gamma_k$ for all users. Left: $\gamma_k=15$ dB, Right: $\gamma_k=25$ dB.