基于广义深度神经网络的多频点统一智能测向方法  

Multi-frequency Unified Direction Finding Method Based On Generalized Deep Neural Network

作　　者：黄凯 尤明懿[1,2] 陆安南 楼财义[1,2] HUANG Kai;YOU Ming-yi;LU An-nan;LOU Cai-yi(Science and Technology on Communication Information Security Control Laboratory,Jiaxing 314033,China;The 36 th Research Institute of CETC,Jiaxing 314033,China)

机构地区：[1]通信信息控制和安全技术重点实验室,浙江嘉兴314033 [2]中国电子科技集团公司第三十六研究所,浙江嘉兴314033

出　　处：《中国电子科学研究院学报》2022年第12期1158-1164,1172,共8页Journal of China Academy of Electronics and Information Technology

摘　　要：文中提出了一种基于深度神经网络(Deep Neural Networks,DNN)用于多频点统一智能测向方法。提出了二维测向的模型输出,以及一种非均匀角度和频率分区的模型训练策略。根据模型输出和分区策略,DNN体系结构由一系列分类网络和角度预测的回归网络组成。随后对智能测向模型进行实测数值分析,给出存在各类误差的非理想接收场景下的二维测向性能,并且对于训练集未包含的频点样本,模型具备较好的测向泛化能力。This paper develops a deep neural network(DNN)based direction finding approach for 2-D direction finding in multiple frequencies.The approach is frequency-unified and is capable of directionof-arrival(DoA)prediction for multiple frequencies.A model output is proposed for 2-D direction finding.An inhomogeneous angle and frequency partition strategy is suggested for efficient implementation.Following the model output and the partition strategy,the general DNN architecture consists of a series of classification networks for angle zoom partition and regression networks for DoA prediction.Numerical investigations are included to evaluate the performance of the proposed approach and show desirable performance for both ideal and nonideal scenarios.The added value of treating samples of different frequencies in a unified manner is also shown.

关 键 词：深度神经网络 测向 非理想场景 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]