# Fast Near-Field Beam Training for Extremely Large-Scale Array

arXiv: https://arxiv.org/abs/2209.14798 · HTML: https://mlsanity.com/paper/2209.14798/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Sep 2022.

## Built on by (most cited first, 1 of 1)

- [Near-field Beam training for Extremely Large-scale MIMO Based on Deep Learning](https://mlsanity.com/paper/2406.03249/) (2024)

## Abstract (arXiv)

In this letter, we study efficient near-field beam training design for the extremely large-scale array (XL-array) communication systems. Compared with the conventional far-field beam training method that searches for the best beam direction only, the near-field beam training is more challenging since it requires a beam search over both the angular and distance domains due to the spherical wavefront propagation model. To reduce the near-field beam-training overhead based on the two-dimensional exhaustive search, we propose in this letter a new two-phase beam training method that decomposes the two-dimensional search into two sequential phases. Specifically, in the first phase, the candidate angles of the user is determined by a new method based on the conventional far-field codebook and angle-domain beam sweeping. Then, a customized polar-domain codebook is employed in the second phase to find the best effective distance of the user given the shortlisted candidate angles. Numerical results show that our proposed two-phase beam training method significantly reduces the training overhead of the exhaustive search and yet achieves comparable beamforming performance for data transmission.