跳到主要导航 跳到搜索 跳到主要内容

Coding programmable metasurfaces based on deep learning techniques

  • Tsinghua University

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

In this paper, we propose a coding scheme based on deep learning techniques for complex beam forming with programmable metasurfaces. A fully convolutional neural network is carefully designed that can 'learn' the physics of beam forming from computed data, and make online prediction of the coding matrices. Both numerical and experimental results show that the network can compute coding matrices fulfilling the input requirement in less than one millisecond. This scheme may open a door for real-time complex beam steering with programmable metasurfaces and digital phased arrays.

源语言英语
主期刊名2019 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2019 - Proceedings
出版商Institute of Electrical and Electronics Engineers Inc.
245-246
页数2
ISBN(电子版)9781728106922
DOI
出版状态已出版 - 7月 2019
已对外发布
活动2019 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2019 - Atlanta, 美国
期限: 7 7月 201912 7月 2019

出版系列

姓名2019 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2019 - Proceedings

会议

会议2019 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting, APSURSI 2019
国家/地区美国
Atlanta
时期7/07/1912/07/19

指纹

探究 'Coding programmable metasurfaces based on deep learning techniques' 的科研主题。它们共同构成独一无二的指纹。

引用此