Joint Microstrip Selection and Beamforming Design for MmWave Systems with Dynamic Metasurface Antennas

Wei Huang, Haiyang Zhang, Nir Shlezinger, Yonina C. EldarICASSP 2023 · DOI

TL;DR

This paper tackles the challenge of efficiently using limited radio frequency (RF) chains in millimeter-wave (mmWave) communication systems that employ dynamic metasurface antennas (DMAs). They propose a novel joint design approach that simultaneously optimizes which antenna elements (microstrips) are activated and how signals are steered (beamforming), achieving performance close to systems with many more RF chains while significantly reducing cost and power consumption.

What Problem Does This Paper Address?

Future wireless systems, particularly those operating in the millimeter-wave (mmWave) bands, promise massive data rates due to the availability of large bandwidths. To achieve this, they require large antenna arrays. However, implementing these arrays with traditional hardware is prohibitively expensive and power-hungry, primarily due to the cost of radio frequency (RF) chains. Dynamic metasurface antennas (DMAs) offer a promising alternative, inherently providing analog beamforming capabilities with lower power consumption. DMAs consist of multiple microstrips, each containing many antenna elements. The challenge arises when the number of available RF chains is limited, which is common in practical systems. This limitation necessitates a strategy to select which microstrips to activate and how to design the beamforming to compensate for the reduced hardware. Existing approaches often involve trade-offs between performance, cost, and complexity. This paper aims to bridge this gap by proposing an integrated design that jointly optimizes microstrip selection and beamforming for mmWave DMAs with a limited number of RF chains.

Methods

The authors propose an iterative, alternating optimization algorithm to solve the joint microstrip selection and beamforming problem. The core idea is to break down the complex, non-convex optimization problem into smaller, manageable subproblems that can be solved efficiently.

  1. Equivalent Channel Modeling: First, they explicitly characterize the equivalent mmWave channel model for DMAs. This model accounts for both the propagation characteristics within the DMA's microstrips and the external wireless channel.
  2. Joint Optimization Problem: They formulate an optimization problem to maximize the signal-to-noise ratio (SNR) subject to constraints on the total transmit power and the number of active RF chains (which dictates the number of active microstrips).
  3. Alternating Optimization: The algorithm iteratively optimizes two key components:
  4. Microstrip Selection and Digital Beamforming (Optimizing w): Given the DMA weights (analog beamforming), this subproblem selects the optimal set of microstrips to activate and designs the digital beamforming vector. This is solved using a channel-gain based selection strategy, effectively picking the microstrips with the strongest channel gains.
  5. DMA Weights (Optimizing G): Given the selected microstrips and the digital beamforming vector, this subproblem optimizes the analog beamforming weights for the active microstrips. This is tackled using a coordinate ascent method, leveraging a transformation to simplify the Lorentz constraint inherent to DMA phase weights.

This alternating approach allows them to converge to a locally optimal solution for the overall problem.

Datasets

Not discussed in the paper. The evaluation is performed using numerical simulations with parameters representative of mmWave systems and DMA characteristics.

Key Insights and Ideas

  • DMA's Unique Channel: The paper highlights that DMAs introduce a unique channel characteristic where the signal's propagation is affected by both the wireless environment and the internal structure of the DMA's microstrips. This necessitates a specific channel model.
  • Synergy of Selection and Beamforming: The most significant insight is that jointly optimizing microstrip selection and beamforming yields superior results compared to treating these as separate problems. This synergy allows the system to intelligently leverage limited RF chains.
  • Low-Complexity Solution: The proposed alternating optimization algorithm is designed to be computationally efficient, making it suitable for practical implementation, potentially even online.
  • Approaching Full-Digital Performance: The results demonstrate that by carefully selecting microstrips and optimizing beamforming, DMA systems with significantly reduced RF chains can achieve performance very close to that of systems with unconstrained RF chains.

Key Results

The proposed joint microstrip selection and beamforming scheme demonstrates significant advantages:

  • RF Chain Reduction: The scheme can reduce the number of required RF chains by up to 75% compared to systems without RF chain reduction.
  • Performance Preservation: Despite the substantial reduction in RF chains, the achievable spectrum efficiency is shown to be comparable to that of DMAs without RF chain reduction, and significantly better than other hybrid architectures with similar RF chain constraints.
  • Cost and Power Savings: By reducing the number of RF chains and associated analog circuitry, the proposed method offers substantial cost and power savings.

Limitations and Future Work

  • Local Optimality: The proposed algorithm guarantees convergence to a local optimum, not necessarily the global optimum, due to the non-convex nature of the original problem.
  • Narrowband Assumption: The current work focuses on narrowband signaling. The authors suggest extending the approach to wideband signals, where DMAs offer additional flexibility for frequency-selective analog processing.
  • Hybrid Architectures: The proposed framework can be extended to other hybrid architectures beyond DMAs, such as those using vector modulators or phase shifter networks.
  • Time-Varying Channels: For rapidly changing channels, the authors propose exploring model-based deep learning techniques to accelerate the optimization process.

Summary updated

BibTeX
@inproceedings{huang2023joint,
  author = {Wei Huang and Haiyang Zhang and Nir Shlezinger and Yonina C. Eldar},
  title = {{Joint Microstrip Selection and Beamforming Design for MmWave Systems with Dynamic Metasurface Antennas}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2023},
  doi = {10.1109/icassp49357.2023.10096260},
}
Figure