Dimensionality Reduced Antenna Array for Beamforming/steering

Shiyi Xia, Mingyang Zhao, Ma Qian, Xunnan Zhang, Ling Yang, Yazhi Pi, Hyun-Chul Chung, Ad C. F. Reniers, A.m.j. Koonen, Zizheng CaoarXiv 2022 · DOI · arXiv

Abstract

Beamforming makes possible a focused communication method. It is extensively employed in many disciplines involving electromagnetic waves, including arrayed ultrasonic, optical, and high-speed wireless communication. Conventional beam steering often requires the addition of separate active amplitude phase control units after each radiating element. The high power consumption and complexity of large-scale phased arrays can be overcome by reducing the number of active controllers, pushing beamforming into satellite communications and deep space exploration. Here, we suggest a brand-new design for a phased array antenna with a dimension reduced cascaded angle offset (DRCAO-PAA). Furthermore, the suggested DRCAO-PAA was compressed by using the concept of singular value deposition. To pave the way for practical application the particle swarm optimization algorithm and deep neural network Transformer were adopted. Based on this theoretical framework, an experimental board was built to verify the theory. Finally, the 16/8/4 -array beam steering was demonstrated by using 4/3/2 active controllers, respectively.

Introduction

Phased array antennas are the backbone of modern high-speed communication, radar, and satellite systems. They work by "beamforming"—precisely controlling the phase and amplitude of signals at each individual antenna element to steer a wireless beam without moving parts. However, the traditional approach is hardware-heavy: a 1-by-N array typically requires N-1 phase shifters. As we move toward massive arrays for 5G/6G and deep space exploration, this leads to prohibitive power consumption, high costs, and extreme system complexity.

The challenge addressed here is whether high-precision beam steering can be maintained while significantly reducing the number of active electronic controllers.

Conceptual overview of the DRCAO-PAA system and matrix-based dimensionality reduction.

Methods

The proposed framework, DRCAO-PAA (Dimension Reduced Cascaded Angle Offset Phased Array Antenna), treats beam steering as a mathematical matrix problem. The methodology involves four key stages:

  1. Matrix Compression via SVD: The steering directions are represented as a coefficient matrix. Using Singular Value Decomposition (SVD), the researchers identified that this matrix is sparse and "low-rank," allowing a few "basis" vectors to represent almost all steering directions.
  2. Linear Independent Vector Group (LIVG): Instead of controlling every element, a small group of base vectors (LIVG) is selected. Desired beam directions are synthesized as linear combinations of these few vectors.
  3. AI Optimization: To find optimal basis vectors, Particle Swarm Optimization (PSO) was used to minimize the difference between ideal and reconstructed beams.
  4. Transformer Model: For real-time application, a Deep Learning Transformer was trained to quickly generate the optimal LIVG for any given design parameters, bypassing the computationally expensive PSO process.

Results and Discussion

The researchers generated a comprehensive synthetic dataset, traversing scanning angles from 5° to 85° and varying antenna element numbers from 4 to 16, to train the Transformer model. The model achieved 89.68% cosine similarity and a low Mean Absolute Error (MAE) of 0.0748 compared to ground-truth PSO results.

Simulated beam steering performance showing amplitude across various scan angles and steps.

Key results include:

  • Hardware Reduction: Simulations of a 128-element array showed that 127 phase shifters could be replaced by just 16 active units with a beam pointing error of only 1.3°.
  • Experimental Validation: A 28 GHz mm-wave prototype demonstrated an 8-element array steered by 3 controllers (62.5% reduction) and a 16-element array steered by 4 controllers (75% reduction).
  • Precision: The 16-element experimental board achieved a 30° scan range with a beam pointing accuracy loss of less than 1.73°.

Conclusion

The DRCAO-PAA design offers a significant reduction in hardware complexity for phased arrays. While dimensionality reduction introduces a manageable "performance penalty" in terms of gain or accuracy, the trade-off is highly beneficial for large-scale deployments. Future work will focus on scaling this approach to massive 2D planar arrays and further optimizing side lobe levels in complex environments.

Summary updated

Abstract

Beamforming makes possible a focused communication method. It is extensively employed in many disciplines involving electromagnetic waves, including arrayed ultrasonic, optical, and high-speed wireless communication. Conventional beam steering often requires the addition of separate active amplitude phase control units after each radiating element. The high power consumption and complexity of large-scale phased arrays can be overcome by reducing the number of active controllers, pushing beamforming into satellite communications and deep space exploration. Here, we suggest a brand-new design for a phased array antenna with a dimension reduced cascaded angle offset (DRCAO-PAA). Furthermore, the suggested DRCAO-PAA was compressed by using the concept of singular value deposition. To pave the way for practical application the particle swarm optimization algorithm and deep neural network Transformer were adopted. Based on this theoretical framework, an experimental board was built to verify the theory. Finally, the 16/8/4 -array beam steering was demonstrated by using 4/3/2 active controllers, respectively.
BibTeX
@misc{xia2022dimensionality,
  author = {Shiyi Xia and Mingyang Zhao and Ma Qian and Xunnan Zhang and Ling Yang and Yazhi Pi and Hyun-Chul Chung and Ad C. F. Reniers and A.M.J. Koonen and Zizheng Cao},
  title = {{Dimensionality Reduced Antenna Array for Beamforming/steering}},
  year = {2022},
  eprint = {2210.16197},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2210.16197},
}
Figure