朱士涛,,
年毅恒,
田春明,
张安学
西安交通大学信息与通信工程学院多功能材料与结构教育部重点实验室 西安 710049
基金项目:国家自然科学基金(62071371, 61801368, 61801366),超高速电路设计与电磁兼容教育部重点实验室(LHJJ/2020-04),雷达信号处理国防科技重点实验室
详细信息
作者简介:周宁宁(1996–),女,河南周口人,西安交通大学信息与通信工程学院在读硕士研究生。主要研究方向为基于OAM的关联成像算法、传输超表面产生OAM
朱士涛(1980–),男,河北沧州人,博士,西安交通大学信息与通信工程学院副研究员,硕士生导师。主要研究方向为新型雷达信号处理方法、人工智能成像算法、微波关联成像、超材料孔径天线及微波量子雷达
年毅恒(1995–),男,安徽蚌埠人,硕士,西安交通大学信息与通信工程学院在读博士研究生。主要研究方向为微波关联成像、雷达信号处理
田春明(1974–),男,吉林通化人,博士,西安交通大学信息与通信工程学院讲师,硕士生导师。主要研究方向为电磁场与微波技术、计算电磁学、高功率微波技术、电磁辐射和散射等
张安学(1972–),男,河南安阳人,博士,西安交通大学电磁与信息技术研究所教授,博士生导师。主要研究方向为新型天线与分集技术、移动通信微波射频技术、智能雷达信号处理、多天线通信系统与阵列信号处理、微波测试理论与系统设计等
通讯作者:朱士涛 shitaozhu@xjtu.edu.cn
责任主编:沙威 Corresponding Editor: SHA Wei中图分类号:TN95
计量
文章访问数:472
HTML全文浏览量:114
PDF下载量:125
被引次数:0
出版历程
收稿日期:2021-04-30
修回日期:2021-07-09
网络出版日期:2021-07-20
刊出日期:2021-10-28
An Intelligent Target Feature Recognition Method Based on Multi-mode OAM Beams
ZHOU Ningning,ZHU Shitao,,
NIAN Yiheng,
TIAN Chunming,
ZHANG Anxue
School of Information and Communication Engineering, Xi’an Jiaotong University, Key Laboratory of Multifunctional Materials and Structures, Ministry of Education, Xi’an 710049, China
Funds:The National Natural Science Foundation of China (62071371, 61801368, 61801366), The Key Laboratory of High-Speed Circuit Design and EMC Ministry of Education (LHJJ/2020-04), The National Key Lab of Radar Signal Processing
More Information
Corresponding author:ZHU Shitao, shitaozhu@xjtu.edu.cn
摘要
摘要:高效率目标探测需要借助探测信号的低相关性空间调制,调制数量大且具有时间独立性。携带轨道角动量(OAM)的涡旋波束具有无穷多种模态且不同模态之间相互正交,借助强色散材料可以实现频率域的多模态OAM波束产生。该文首先对OAM的传播特性进行推导,给出了符合探测需求的多模态OAM波束源特征;在此基础上,研究了不同模态的OAM波束在3种不同应用场景下目标反射回波信号特性,采用卷积神经网络对不同反射场景下的数据特征进行提取,实现了对未知场景的判断及场景内的目标识别,并进行了抗噪性能分析。实验结果表明:理想状态下,网络对目标场景判断的准确率可达
![](https://radars.ie.ac.cn/article/doi/10.12000/R21056_M1.jpg)
![](https://radars.ie.ac.cn/article/doi/10.12000/R21056_M2.jpg)
![](https://radars.ie.ac.cn/article/doi/10.12000/R21056_M3.jpg)
关键词:空间调制/
轨道角动量(OAM)/
多模态涡旋电磁波/
卷积神经网络/
特征提取/
目标识别
Abstract:Target detection based on space modulation requires a large number of test modes with space-time independence. The Orbital Angular Momentum (OAM) beams are orthogonal to each other and have infinite modes. Due to the strong dispersive materials, multi-mode OAM beams with the same scattering angle can be generated in the frequency domain. In this manuscript, the propagation characteristics of multi-mode OAM beams are analyzed, which can be utilized to improve detection efficiency. The echoes from the target illuminated by the multi-mode OAM beams are then investigated in three different application scenarios. A convolution neural network is employed to extract the relationship between the echo data and the target image based on prior knowledge. The target and the imaging scenarios can be distinguished with a high probability. Finally, the proposed method’s anti-noise performance is analyzed. The experimental results show that in the ideal state, the accuracy of target scene judgment can reach 97.5%. The accuracy of the target location recognition is higher than 80% when the interval between two adjacent targets in a scene is larger than a threshold. The accuracy of the target location recognition in three scenes is greatly reduced when SNR is less than 20 dB, depending on the scene.
Key words:Spatial modulation/
Orbital Angular Momentum (OAM)/
Multi-mode vortex electromagnetic wave/
Convolution Neural Network (CNN)/
Feature extraction/
Target recognition
PDF全文下载地址:
https://plugin.sowise.cn/viewpdf/198_d04f5f8d-7f73-42ee-91a9-47c4e62bc487_R21056