Few-shot electromagnetic signal classification: A data union augmentation method

Deep learning has been fully verified and accepted in the field of electromagnetic signal classification. However, in many specific scenarios, such as radio resource management for aircraft communications, labeled data are difficult to obtain, which makes the best deep learning methods at present seem almost powerless, because these methods need a large amount of labeled data for training. When the training dataset is small, it is highly possible to fall into overfitting, which causes performance degradation of the deep neural network. For few-shot electromagnetic signal classification, data augmentation is one of the most intuitive countermeasures. In this work, a generative adversarial network based on the data augmentation method is proposed to achieve better classification performance for electromagnetic signals. Based on the similarity principle, a screening mechanism is established to obtain high-quality generated signals. Then, a data union augmentation algorithm is designed by introducing spatiotemporally flipped shapes of the signal. To verify the effectiveness of the proposed data augmentation algorithm, experiments are conducted on the RADIOML 2016.04C dataset and real-world ACARS dataset. The experimental results show that the proposed method significantly improves the performance of few-shot electromagnetic signal classification.     

通信信号调制识别综述

随着频谱信息资源愈发紧张,信号的调制处理方式也愈发多样化.由于通信信号的调制识别广泛应用于民用和军用领域,所以,对调制方式进行识别的研究具有极其重要的意义与战术价值.首先,根据大量文献,对基于最大似然函数和特征提取的经典调制识别方法进行了系统地梳理;其次,从常用数据集、深度神经网络结构和现有识别算法等方面,着重介绍了深...     

基于深度学习的ADS-B异常数据检测模型

广播式自动相关监视（ADS-B）是下一代空中交通运输系统的重要组成部分，是新航行系统中非常重要的通信和监视技术，但其协议没有提供相关的信息认证和数据加密，因此极其容易受到欺骗干扰的影响。针对ADS-B报文数据特点，采用深度学习的seq2seq模型对ADS-B报文数据进行重构，通过重构误差来检测异常，并对数据进行特征扩展，使模型能更好的捕捉数据的时间依赖性。实验结果表明，所采用的方法优于传统的机器学习方法，且在数据特征扩展后，模型检测效果提升。相比于现有的欺骗干扰检测方法，该方法不需要改变ADS-B系统的协议，也不需要额外的节点或传感器参与，具有一定的适应性和灵活性。     

基于BiGRU-SVDD的ADS-B异常数据检测模型

广播式自动相关监视（ADS-B）作为新一代空管监视技术，由于采用明文方式广播发送数据，因而存在易遭受网络攻击的安全问题。为了准确检测ADS-B数据攻击行为，在充分考虑时间相关性的基础上，提出了针对ADS-B数据的异常数据检测模型。首先利用双向门控循环单元（BiGRU）神经网络预测ADS-B数据，得到了ADS-B数据预测值。再将预测值和实际值作差，将差值放入支持向量数据描述（SVDD）训练，得到了能检测ADS-B异常数据的超球体分类器。并且，选择了合适的滑动窗口，在保证异常检测准确率的同时，缩短BiGRU神经网络的训练时长。实验结果表明，BiGRU-SVDD模型能检测出随机位置偏移攻击、高度偏差攻击、重放攻击、拒绝服务（DOS）等攻击下的ADS-B异常数据。并且，与其他机器学习和深度学习方法相比，BiGRU-SVDD异常检测模型的准确率更佳，适应性更优。     

深空探测器无线电干涉测量开源数据处理系统构建及试验

为更好地开展深空无线电干涉测量试验,基于DiFX(Distributed FX,分布式FX型相关处理)、SPICE(Spacecraft Planet Instrument Camera-matrix Events,航天器行星仪器照相机矩阵事件)、HOPS(Haystack Observatory Postprocessing System,Haystack克天文台后处理系统)及AIPS(Astronomical Image Processing System,天文图像处理系统)等开源软件,搭建了一套可用于深空探测器信号相关处理的无线电干涉测量数据处理系统。首先介绍了该系统的整体架构及组成,然后介绍了系统各组成模块的功能及工作原理,最后利用该系统对嫦娥三号着陆器开展了观测试验,并对观测数据进行了处理。试验结果表明,该系统能成功实现深空航天器信号的相关处理、条纹拟合,以及时延解算等功能,有望为我国嫦娥五号任务及火星探测任务数据处理提供支持。     

ADS-B统计数据的位置导航不确定类别质量分析

为实现广播式自动相关监视(ADS-B)航迹与雷达航迹的高质量融合,必须对ADS-B航迹和雷达航迹数据质量进行分析。在分析实际ADS-B航迹数据的基础上,提出以航路为参考基准验证ADS-B航迹数据精度及完好性的方法。根据ADS-B报文中的位置导航不确定类别(NUCp)值,统计航迹点与航路中心线的水平偏差并将其作为位置实际误差,运用分位数-分位数图法对位置实际误差进行正态分布检验,得出置信水平为99%的实际误差置信区间,即在不同NUCp值下可信赖的位置误差范围。实验结果表明,在采用加权平均法进行航迹融合时,与传统权重分配相比,对于NUCp值为5或6的ADS-B航迹数据可增大权重,而对于NUCp值为7的数据则应减小权重。     

基于离散小波变换的近程信号自动识别

进行信号识别的关键在于利用有限信号数据提取出有效的信号特征,然后根据这些特征作出判决。文中阐述了利用离散小波变换对近程信号进行分解处理,并结合具体应用背景,提取小波分解的细节作为信号识别的特征参量。仿真实验表明,基于小波分解的特征提取过程对噪声有较好的抑制能力,小波变换的应用能够有效提高对近程信号识别的能力。     

ADS-B监视数据质量分析研究

为了进一步推广ADS-B技术,提高ADS-B数据的质量,通过对ADS-B数据的分析,研发了一种ADS-B数据质量评估系统,该系统可以对CAT021报文进行解析,并对解析后的数据的完好性、漏点率和跳点率进行分析,以及航班轨迹的显示。论文中采用了广州＂大和章＂ADS-B台站数据对其进行了验证,其评估结果可靠有效,对ADS-B数据分析研究有一定的参考意义。     

含噪声的转子碰摩混沌信号分类识别

采用基于竞争学习和聚类分析的学习向量量化 ( LVQ)方法 ,研究转子碰摩混沌响应信号的神经网络分类识别问题 ,给出了相应的理论分析和计算结果 ,着重研究了 LVQ网络在不同噪声时的识别情况。分析结果表明 ,该方法可以实现转子碰摩混沌信号与其它响应信号的分类识别 ,并且具有良好的抗噪性能 ,为转子碰摩混沌信号的分类识别提供了一种较为直接的实时处理方法。     

Blind identification of space–time block codes based on deep learning

Deep Learning(DL) has important applications to both commercial and military communications, such as software-defined radio, cognitive radio and spectrum surveillance. While DL has been intensively studied for modulation recognition, there are very few investigations for blind identification of Space-Time Block Codes(STBCs). This paper proposes a Residual Network(RN)-based model for identifying 6 kinds of STBC signals with a single receiving antenna, including the same length of coding matrix. I...     

A visionary look at aviation surveillance systems

The FAA's future aviation surveillance systems fall into four categories: Automatic Dependent Surveillance (ADS) will be used in the oceanic environment; ADS-Broadcast (ADS-B) will be used in the domestic en route environment; ADS-B will be used with a secondary radar backup in the terminal area; and ADS-B will be used with primary radar backup within the Airport Surface Traffic Automation (ASTA) system on the airport's surface environment. Two other systems introduced in this paper are Cockpit Display of Traffic Information (CDTI) and Traffic Advisory and Collision Avoidance System (TCAS). All these systems will use navigational signals emitted by the Global Positioning System (GPS) constellation of satellites     

考虑噪声抑制的高超声速风洞气动力识别方法

对于风洞试验中全尺寸模型试验的非平稳信号进行载荷辨识仍存在诸多问题。针对全尺度模型试验的非平稳信号载荷辨识提出了一种基于深度残差收缩网络（DRSN）深度学习技术的智能载荷辨识方法,该方法通过深度学习提取测力系统输出数据中的气动力、惯性力和噪声等特征,通过注意力机制对每组数据进行获取阈值,再通过软阈值函数对特征进行滤波降噪,有效辨识出测力系统响应信号中的惯性力分量并进行剔除,实现气动力载荷辨识。在测试验证中,均值法的辨识精度为85%以上,DRSN模型的辨识精度为94%以上,证明DRSN模型能有效降低噪声和惯性力对于载荷辨识的干扰,用于非平稳信号的载荷辨识具有精度高、可靠性好等特点。     

基于深度卷积网络的SAR图像目标检测识别

在SAR图像解译应用领域,目标的自动检测与识别一直是该领域的研究重点和热点,也是该领域的研究难点。针对SAR图像的目标检测与识别方法一般由滤波、分割、特征提取和目标识别等多个相互独立的步骤组成。复杂的流程不仅限制了SAR图像目标检测识别的效率,多步骤处理也使模型的整体优化难以进行,进而制约了目标检测识别的精度。采用近几年在计算机视觉领域表现突出的深度学习方法来处理SAR图像的目标检测识别问题,通过使用CNN、Fast RCNN以及Faster RCNN等模型对MSTAR SAR公开数据集进行目标识别及目标检测实验,验证了卷积神经网络在SAR图像目标识别领域的有效性及高效性,为后续该领域的进一步研究应用奠定了基础。     

Study of ADS-B Data Evaluation

In western China, the terrain and meteorological conditions are so complex that it is not suitable to construct new radar stations. Automatic dependent surveillance-broadcast (ADS-B) is a totally new surveillance method, so before practically applied and operated in China, abundant tests and evaluations are necessary to validate the performance of ADS-B and guarantee the operational security. During the flight tests, we collect the data of radar, ADS-B and high accuracy position and compare the performance of ADS-B with radar based on high accuracy position. To solve the asynchronous problem among radar data, ADS-B data and real-time kinematic (RTK) data caused by different update rates, this article proposes the technique of synchronizing multi-surveillance data by extrapolating from the data of low update rate to high update rate according to velocity and heading. Meanwhile, because radar data, ADS-B data and RTK data are expressed in different coordinates and cannot be compared each other, this article provides a method to unifying the coordinates of multi-surveillance data. By the analysis and evaluation, we can conclude that the performance of ADS-B is better than radar.     

基于软件无线电的机会信号实时盲分离系统

在复杂的电磁环境中,往往存在多种频率交叠的机会信号,为利用机会信号进行定位增加了难度。针对这个问题,提出了一种基于软件无线电的机会信号实时盲分离系统。系统的硬件部分是在FPGA中实现等变自适应盲分离算法的,实现了AD采集后的复数域混合信号的分离。分离后的信号通过USB3.0数据线上传到上位机中进行存储、识别和实时显示。算法的FPGA实现采用流水线架构和浮点运算设计来提高信号的实时处理能力和数据计算精度。实测结果表明,该系统能够实时分离多路复数域混合机会信号。     

EBAA：An efficient broadcast authentication scheme for ADS-B communication based on IBS-MR

Automatic dependent surveillance-broadcast （ADS-B） systems can broadcast satellitebased aircraft position, identification, etc., periodically, and are now on track to replace radar to become the backbone of next-generation air traffic management （ATM） systems. However, ADS-B systems suffer severe cyber-security problems due to the broadcast-type data link and the lack of designed-in security measures. Especially, since ADS-B messages are unauthenticated, it is easy to insert fake aircraft into a system via spoofing or insertion of false messages. Unfortu- nately, the authentication for ADS-B messages has not yet been well studied. In this paper, based on identity-based signature with message recovery （IBS-MR）, an efficient broadcast authentication scheme for ADS-B messages is proposed. The security analysis demonstrates that the scheme can achieve authenticity and integrity of ADS-B broadcast messages, as well as adaptive evolution of broadcasters＇ private keys. The performance evaluation shows that the scheme is computationally efficient for typical avionics devices with limited resources. Furthermore, the scheme achieves low communication overhead since broadcast messages can be recovered from signatures, and thus it is suitable for low-bandwidth ADS-B data link.     

Detection Probabilities for Log-Normally Distributed Signals

The amplitude and power of a large family of radio signals are observed to have log-normal probability density functions. Among these are signals propagated through random inhomogeneous media, a notable example being low frequency atmospheric radio noise. Of greater importance are certain radar targets that have been observed to have essentially log-normal density functions. Both ships and space vehicles may fall into this category. Curves of probability of detection vs. signal-to-noise ratio for the case of log-normal signals in Gaussian noise have been computed and are presented in this paper. The curves apply for square-law detection with varying degrees of postdetection linear integration. Both fully correlated and completely uncorrelated fluctuating signals are considered. It is shown that for log-normal signal distributions having large variances, the probability of detection differs significantly from that obtained using curves based on an assumed Rayleigh signal distribution.     

基于相频约束的航天器定时技术

双向相干测距测速体制是最常用的航天器测量体制,是一种闭环体制.与之相比较,开环测距测速可以在信号更微弱的情况下获取到观测量,在深空任务中有着独特的意义.然而,开环测距也面临棘手的技术难题——高精度航天器定时技术.针对该难题,在保持现有航天器信号体制的基础上,阐述了基于遥测信号和测距侧音信号（差分单向测距DOR信号可看作侧音信号）间相频约束的航天器定时原理,研究了侧音频率最优化设计方法,给出了“器上发射测站接收”的初步实现方案,为后续开展工程应用奠定了基础.     

Synchronized perturbation elimination and DOA estimation via signal selection mechanism and parallel deep capsule networks in multipath environment

State-of-the-art model-driven Direction-Of-Arrival (DOA) estimation methods for multipath signals face great challenges in practical application because of the dependence on the precise multipath model. In this paper, we introduce a framework, based on deep learning, for synchronizing perturbation auto-elimination with effective DOA estimation in multipath environment. Firstly, a signal selection mechanism is introduced to roughly locate specific signals to spatial subregion via frequency domain filters and compressive sensing-based method. Then, we set the mean of the correlation matrix’s row vectors as the input feature to construct the spatial spectrum by the corresponding single network within the parallel deep capsule networks. The proposed method enhances the generalization capability to untrained scenarios and the adaptability to non-ideal conditions, e.g., lower SNRs, smaller snapshots, unknown reflection coefficients and perturbational steering vectors, which make up for the defects of the previous model-driven methods. Simulations are carried out to demonstrate the superiority of the proposed method.     

机载ADS-B技术现状与发展趋势

随着广播式自动相关监视( ADS-B )技术的发展与广泛应用,国外涌现出许多 ADS-B 设备研制商,开发出多种应用于不同平台、具有不同功能构型的 ADS灢B 产品。从搭载平台与功能构型两个视角出发,对国内外机载 ADS-B 产品进行分类;分析航线飞机、通航飞机以及无人机载 ADS灢B 产品的功能与技术特点,并针对当前 ADS-B 系统面临航空器监视应用开发、安全性、密集干扰环境下接收增强、1090ES 链路扩容等难点问题,分析讨论技术途径并提出发展建议。