基于权重不可知神经网络的旋翼无人机检测  

Micro-motion Feature Detection of Rotary-wing UAV Based on Weight Agnostic Neural Network

作　　者：谢跃雷[1,3] 刘信 梁文斌 XIE Yuelei;LIU Xin;LIANG Wenbin(School of Information and Communication,Guilin University of Electronic Technology,Guilin 541004,China;School of Information Science,Guilin University of Electronic Technology,Guilin 541004,China;Guangxi Key Laboratory of Wireless Wideband Communication and Signal Processing,Guilin 541004,China)

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004 [2]桂林电子科技大学信息科技学院,广西桂林541004 [3]广西无线宽带通信与信号处理重点实验室,广西桂林541004

出　　处：《电讯技术》2022年第1期46-53,共8页Telecommunication Engineering

基　　金：国家自然科学基金资助项目(6146105);广西科技重大专项资助项目(桂科AA21077008);广西高校中青年教师科研基础能力提升项目(2019KY1035);桂林电子科技大学研究生教育创新计划项目(2020YCXS021)。

摘　　要：针对传统无人机检测方法缺乏智能性和泛化性的问题,提出了一种基于权重不可知神经网络(Weight Agnostic Neural Network,WANN)的无人机微动特征检测方法,以实现探测无人机的目的。推导了旋翼无人机微动模型,详细说明了WANN模型的构建过程。以回波信号的循环谱等高图作为训练、测试数据集进行了仿真,结果表明该方法对噪声有较好的鲁棒性。实测结果也验证了WANN模型能有效提取无人机微动特征,有较好的图像识别能力。In view of the lack of intelligence and generalization of traditional unmanned aerial vehicle(UAV)detection methods,this paper proposes a UAV micro-motion feature detection method based on Weight Agnostic Neural Network(WANN)to achieve the purpose of detecting UAV.The micro-motion model of the rotor UAV is derived and then the construction process of WANN model is explained in detail.By using the cyclic spectral contour map of echo signal as the training and testing data set,simulation is performed.The results show that the proposed method is robust to noise.At the same time,the actual measurement results also verify that WANN model can effectively extract the micro-motion features of UAV and has a good ability of image recognition.

关 键 词：无人机(UAV) 微动特征检测 权重不可知神经网络(WANN) 循环谱等高图 

分 类 号：TN957[电子电信—信号与信息处理] V279[电子电信—信息与通信工程]