Yuan Ye, Wu Shuang, Yang Yong, Yuan Naichang
State Key Laboratory of Complex Electromagnetic Environment Elects on Electronics and Information System, National University of Defense Technology, Changsha, 410073, China.
Innovation Academy for Microsatellites, Chinese Academy of Sciences, Shanghai, 200120, China.
Sci Rep. 2021 Mar 18;11(1):6386. doi: 10.1038/s41598-021-85864-5.
Deep neural networks have shown great performance for direction-of-arrival (DOA) estimation problem, but it is necessary to design some suitable networks to solve the multi-DOA estimation problem. In this paper, we use Khatri-Rao product to increase the degree of freedom of antenna array and obtain the image tensor of covariance matrix, then we propose an improved estimation network to process the tensor. We use the curriculum learning scheme and partial label strategy to develop a CurriculumNet training scheme. The training/validation results shows that the proposed training scheme can increase the generalization of the estimation network and improve the accuracy of network around [Formula: see text]. The estimation performance of the proposed network shows high-resolution results, which can distinguish two adjacent signals with angle difference of [Formula: see text]. Moreover, the proposed estimation network has root mean square estimation error lower than [Formula: see text] when signal noise ratio equals [Formula: see text] and can estimate DOAs precisely by only 8 snapshots, which performs much better than prior deep neural network based estimation methods and can estimate multi-DOA results under hostile estimation environments.
深度神经网络在到达方向（DOA）估计问题上表现出了优异的性能，但有必要设计一些合适的网络来解决多DOA估计问题。在本文中，我们使用Khatri-Rao积来增加天线阵列的自由度并获得协方差矩阵的图像张量，然后我们提出一种改进的估计网络来处理该张量。我们使用课程学习方案和部分标签策略来开发一种课程网络训练方案。训练/验证结果表明，所提出的训练方案可以提高估计网络的泛化能力，并在[公式：见原文]左右提高网络的准确性。所提出网络的估计性能显示出高分辨率结果，能够区分角度差为[公式：见原文]的两个相邻信号。此外，当信噪比等于[公式：见原文]时，所提出的估计网络的均方根估计误差低于[公式：见原文]，并且仅通过8个快照就能精确估计DOA，其性能比现有的基于深度神经网络的估计方法要好得多，并且能够在恶劣的估计环境下估计多DOA结果。
