A Low-Profile Broadband Folded Transmitarray Shared-Aperture Transceiver for Frequency-Diverse Near-Field Sensing Enabled by Convolutional Neural Networks

  • Jinyu Wu
  • , Dingfei Ma
  • , Xinhao Chen
  • , Qingfeng Zhang
  • , Huan Meng
  • , Kin Fai Tong
  • , Baiyang Liu

Research output: Contribution to journalArticlepeer-review

Abstract

This letter introduces a low-profile broadband folded transmitarray shared-aperture transceiver designed for frequencydiverse near-field sensing and real-time gesture recognition. The system uses a random encoded folded transmitarray and a dualpolarized feed to generate random transmit–receive fields through polarization control. The folded array operates from 31 GHz to 46 GHz (corresponding to 38% relative bandwidth) and includes a random-coded transmitarray surface and a polarization conversion reflectarray surface. The folded design integrates separate transmitter and receiver components into a shared-aperture transmitter yielding a compact 15.4λ × 15.4λ × 2.6λ system with a height-to-diameter ratio of 0.16. Experimental results demonstrate the effectiveness of the proposed compact system in a real-time classifying ten gestures with an accuracy of 99.6% using a convolutional neural network.

Original languageEnglish
Pages (from-to)428-432
Number of pages5
JournalIEEE Antennas and Wireless Propagation Letters
Volume25
Issue number1
DOIs
Publication statusPublished - 2026

Keywords

  • Convolutional neural network (CNN)
  • folded transmitarray
  • frequency diversity
  • near-field sensing

Fingerprint

Dive into the research topics of 'A Low-Profile Broadband Folded Transmitarray Shared-Aperture Transceiver for Frequency-Diverse Near-Field Sensing Enabled by Convolutional Neural Networks'. Together they form a unique fingerprint.

Cite this