A lightweight false alarm suppression method in heterogeneous change detection

Overlooking the issue of false alarm suppression in heterogeneous change detection leads to inferior detection per-formance.This paper proposes a method to handle false alarms in heterogeneous change detection.A lightweight network of two channels is bulit based on the combination of convolutional neural network(CNN)and graph convolutional network(GCN).CNNs learn feature difference maps of multitemporal images,and attention modules adaptively fuse CNN-based and graph-based features for different scales.GCNs with a new kernel filter adaptively distinguish between nodes with the same and those with different labels,generating change maps.Experimental evaluation on two datasets validates the efficacy of the pro-posed method in addressing false alarms.

Constant False Alarm Rate Acquisition Algorithm for Compass B1C Signal

In order to improve the sensitivity of the Compass B1C signal acquisition for the receiver,the principle of constant false alarm rate(CFAR)is applied for the B1C pilot channel acquisition to realize the dynamic adjustment of the threshold of acquisition against the carrier to noise ratio.The non-coherent data/pilot combined acquisition algorithm for B1C signal is analyzed to make full use of the power of the B1C signal under the condition of low carrier to noise ratio.On this basis,to improve the acquisition sensitivity of the receiver,the principle of constant false alarm probability is applied for the non-coherent data/pilot combined acquisition algorithm.Theoretical analysis and simulations show that the non-coherent data/pilot combined acquisition algorithm with CFAR improves the B1C signal acquisition sensitivity of the receiver significantly,and achieves a better Receiver Operating Characteristic compared with the traditional acquisition algorithms.

SCADA Data-Based Support Vector Machine for False Alarm Identification for Wind Turbine Management

Maintenance operations have a critical influence on power gen-eration by wind turbines(WT).Advanced algorithms must analyze large volume of data from condition monitoring systems(CMS)to determine the actual working conditions and avoid false alarms.This paper proposes different support vector machine(SVM)algorithms for the prediction and detection of false alarms.K-Fold cross-validation(CV)is applied to evaluate the classification reliability of these algorithms.Supervisory Control and Data Acquisition(SCADA)data from an operating WT are applied to test the proposed approach.The results from the quadratic SVM showed an accuracy rate of 98.6%.Misclassifications from the confusion matrix,alarm log and maintenance records are analyzed to obtain quantitative information and determine if it is a false alarm.The classifier reduces the number of false alarms called misclassifications by 25%.These results demonstrate that the proposed approach presents high reliability and accuracy in false alarm identification.

Performance of Order-statistics Constant-false-alarm-rate Detector with Noncoherent Integration and Its Application to OTH Radar

Noncoherent integration is often ed for approving performance in detection of radar signal. Order-statistics constant false alarm rate (OS-CFAR) detector has some advantages in clutter and multiple target situations. AnOS-CFAN detector with noncoherent integration after Square law envelope detector is presented and an analysis of detection performance for the chi-Square family of Swerling fluctuating targets is given. Its application to the high frequency(HF) ground wave over-the-horizon (OTH) radar is discussed as well.

A Double-threshold Constant False Alarm Rate Detector And Its Performance Analysis

he cell averaging and the order statistics are two typical algorithms for constant false alarm rate detector in radar system. They have different advantages in stationary noise background and fluctuation clutter environment respectively. This paper presents a doublethreshold constant false alarm rate detector constructed on the basis of synthesizing the advantages of the two algorithms above and avioding their disadvantages. The performance of the detector is analyzed, and the simulation result is given.

基于威胁机制-双重深度Q网络的多功能雷达认知干扰决策

针对传统深度Q网络(deep Q network,DQN)在雷达认知干扰决策中容易产生经验遗忘,从而重复执行错误决策的问题,本文提出了一种基于威胁机制双重深度Q网络(threat warning mechanism-double DQN,TW-DDQN)的认知干扰决策方法,该机制包含威胁网络和经验回放2种机制。为了验证算法的有效性,在考虑多功能雷达(multifunctional radar,MFR)工作状态与干扰样式之间的关联性的前提下,搭建了基于认知电子战的仿真环境,分析了雷达与干扰机之间的对抗博弈过程,并且在使用TW-DDQN进行训练的过程中,讨论了威胁半径与威胁步长参数的不同对训练过程的影响。仿真实验结果表明,干扰机通过自主学习成功与雷达进行了长时间的博弈,有80%的概率成功突防,训练效果明显优于传统DQN和优先经验回放DDQN(prioritized experience replay-DDQN,PER-DDQN)。

储罐自动化仪表计量误差在线识别方法研究

针对站场油气储罐计量仪表的检测误差,以现场366座储罐正常计量数据为基础,设置0.5%~2.5%的误差率,构建数据驱动方法数据集,综合考虑误报率和窗口大小(25~125),针对7类算法(WIN-G、WIN-M、WIN-L、RuLSIF、KL-CPD、VAE和LSTM-VAE)进行综合性能评价,用现场检测的误差数据和正常数据进行算法验证,结果表明:在窗口大小100的条件下,LSTM-VAE算法性能最佳,正确报警率高于0.95。

Analysis of the False Alarm Fault in Antenna-servo System of Weather Radar

A false alarm fault frequently appeared in antenna-servo system of the CINRAD/SA weather radar of Shanwei in the second half of 2011, so possible reasons for the false alarm fault were listed firstly using method of exhaustion, and then the main reason was determined using exclusive method. That is, the fault was closely related to the signal transmission channel from the antenna mount to servo system in RDA cabinet. After ex- amining questionable nodes in the transmission channels of the alarm signal, we found that the false alarm fault might result from the interference of a burr in the temperature sensing circuit of the elevation motor. In actual operation, a filter capacitor was connected with the corresponding pin in the upper optical board to screen the interference of a burr, thereby successfully eliminating the false alarm fault in antenna-servo system of the CIN- RAD/SA radar of Shanwei.

多电飞机电气系统BIT虚警分析及解决方案

对多电飞机电气系统机内测试(Built-in Test, BIT) 虚警产生的原因进行了分析和总结, 从多电飞机的电源系统、配电系统、电力作动系统和综合控制系统四个方面阐述电气系统的虚警问题, 并针对这几部分的虚警原因, 从BIT的设计方案、系统仿真建模、硬件及软件等几个方面, 提出了一个电气系统BIT虚警问题的系统解决方案, 为多电飞机电气系统测试性及可靠性的深入研究打下了基础。

Frame Detection Based on Cyclic Autocorrelation and Constant False Alarm Rate in Burst Communication Systems

Frame detection is important in burst communication systems for its contribu- tions in frame synchronization. It locates the information bits in the received data stream at receivers. To realize frame detection in the presence of additive white Gaussian noise （AWGN） and frequency offset, a constant false alarm rate （CFAR） detector is proposed through exploitation of cyclic autocorrelation feature implied in the preamble. The frame detection can be achieved prior to bit timing recovery. The threshold setting is independent of the signal level and noise level by utilizing CFAR method. Mathematical expressions is derived in AWGN channel by considering the probability of false alarm and probability of detection, separately. Given the probability of false alarm, the mathematical relationship between the frame detection performance and EJNo of received signals is established. Ex- perimental results are also presented in accor- dance with analysis.

A Secure Framework for WSN-IoT Using Deep Learning for Enhanced Intrusion Detection

The security of the wireless sensor network-Internet of Things(WSN-IoT)network is more challenging due to its randomness and self-organized nature.Intrusion detection is one of the key methodologies utilized to ensure the security of the network.Conventional intrusion detection mechanisms have issues such as higher misclassification rates,increased model complexity,insignificant feature extraction,increased training time,increased run time complexity,computation overhead,failure to identify new attacks,increased energy consumption,and a variety of other factors that limit the performance of the intrusion system model.In this research a security framework for WSN-IoT,through a deep learning technique is introduced using Modified Fuzzy-Adaptive DenseNet(MF_AdaDenseNet)and is benchmarked with datasets like NSL-KDD,UNSWNB15,CIDDS-001,Edge IIoT,Bot IoT.In this,the optimal feature selection using Capturing Dingo Optimization(CDO)is devised to acquire relevant features by removing redundant features.The proposed MF_AdaDenseNet intrusion detection model offers significant benefits by utilizing optimal feature selection with the CDO algorithm.This results in enhanced Detection Capacity with minimal computation complexity,as well as a reduction in False Alarm Rate(FAR)due to the consideration of classification error in the fitness estimation.As a result,the combined CDO-based feature selection and MF_AdaDenseNet intrusion detection mechanism outperform other state-of-the-art techniques,achieving maximal Detection Capacity,precision,recall,and F-Measure of 99.46%,99.54%,99.91%,and 99.68%,respectively,along with minimal FAR and Mean Absolute Error(MAE)of 0.9%and 0.11.

海杂波测量与建模研究进展

海杂波作为对海雷达目标检测的重要影响因素,其特性的准确认知一直是雷达技术领域的关注热点和难点问题.本文从海杂波测量、数据处理和特性建模三个方面,对中国电波传播研究所在海杂波系统性研究方面的进展进行综述.结合具体试验测量系统和数据,介绍了不同平台的海杂波测量技术和数据处理方法,从传统经验建模和基于深度学习的人工智能建模两个角度,对海杂波特性建模进行了重点阐述.最后,从测量、建模和理论研究等方面对海杂波研究未来发展方向进行了展望.

基于ML估计的高动态GNSS信号快速捕获检测方法

针对高动态环境下GNSS因频域带宽增加导致捕获难度增大的问题,分析了接收端数字中频采样信号的传输特性及复基带信号经FFT模块处理后的相关峰的检测,提出了基于极大似然(ML)估计的高动态GNSS信号快速捕获检测方法。首先,根据随机信号的统计理论建立二元假设检验条件,构建了奈曼-皮尔逊准则下的GNSS信号捕获判决门限模型;其次,通过判决量的统计特性对等效高斯白噪声方差进行ML估计,根据其估计值计算捕获判决门限,其中通过虚警率的量化放大处理,解决了判决量样本值的增加带来的估计偏差问题;最后,对不同高动态条件下北斗B3I信号进行了捕获检测仿真实验。结果表明采用ML估计方法确定捕获判决门限从而提高高动态GNSS信号捕获的检测方法对高动态适应范围较宽,其频移捕获精度与SINS信息辅助捕获相当,比序贯检测算法提高约28%以上,相同条件下具有更快的平均捕获检测速度。

Solution of false alarm and slow response in flame detector

A new flame detector with one ultraviolet and two infrared detectors is designed. The ultraviolet detector is of rapid response(≤10 μs) while the two infrared detectors usually have a response time of more than 5 ms. The ultraviolet detector is applied to deal with the flame of large scales. When facing the flame of mid or small scales, the three detectors cooperate. Employing the high-order derivatives of the sample data of the infrared circuits to improve the sensitivity, the response speed is greatly improved. The data of the temperature sensor is used to adjust circuit parameters in real time, thus reducing the effect of temperature drift. The flame detectors are tested at different distances and the response time is as rapid as 0.65 ms. The test results show that the new flame detector has the characteristics of high speed and a low rate of false alarms.

基于杂波拖尾分布的雷达无人机检测性能分析

固定翼无人机(unmanned aerial vehicle,UAV)给雷达低空监视提出了严峻挑战。分析雷达对固定翼UAV的检测性能,可为雷达UAV检测能力评估和技术升级提供重要参考。本文结合雷达探测低空固定翼UAV外场实测数据,首先分析了低空固定翼UAV雷达接收信号幅度统计分布,采用多项式对杂波拖尾导致的虚警概率进行拟合建模;然后,根据虚警概率分布得到雷达检测门限;进而根据UAV回波+杂波幅度分布理论推导得到雷达检测概率;最后,将理论分析性能与传统性能分析结果、雷达实际检测性能进行对比。结果表明,采用多项式对杂波拖尾导致的虚警概率进行单独建模,由此获得的雷达检测门限精度更高,从而使雷达UAV检测性能分析结果较传统性能分析结果更准确。

基于WGAN-GP-CNN的海面小目标检测

针对传统基于统计理论的海面小目标检测方法在复杂海面环境中性能不高的问题,该文提出了一种改进的检测方法。首先通过分析海杂波和目标回波的特征,将检测问题转化为特征空间的分类任务。鉴于海面小目标样本数量有限,存在样本不平衡的问题,该文引入了一种基于梯度惩罚的沃瑟斯坦生成对抗网络(Wasserstein Generative Adversarial Network with Gradient Penalty,WGAN-GP)来增强目标数据,从而在数量上平衡目标样本与海杂波样本。同时,对原始WGAN-GP网络的损失函数进行了改进,引入相位损失以确保生成数据能够反映真实数据的相位信息。基于这些数据,进一步提取了生成目标和海杂波的高维特征,并将其送入卷积神经网络(Convolutional Neural Network,CNN)进行训练。为了应对高维特征空间中虚警概率难以控制的问题,对CNN算法进行了改进,通过设置Softmax分类器的阈值,实现了虚警概率可控。最后,借助公开的IPIX雷达数据集进行实验验证,所提的WGAN-GP-CNN检测器在积累时间为1.024 s,虚警概率为0.001时,平均检测概率达到0.8683,具有良好的检测效果。

基于实时能量估计的自适应码捕获

提出基于实时能量估计的码捕获自适应门限调整方法,给出了任意L次独立观测时虚警概率和检测概率的表达式。理论分析和计算机仿真结果表明:此方法的虚警概率恒定,比MLAP和0SAP等恒虚警检验方法的性能优越,且适用于多径误落信道条件下的RAKE接收。

Support-Vector-Machine-Based False Alarm Filter of Mechatronic Built-in Test

Diagnosing intermittent fault is an important approach to reduce built-in test（BIT） false alarms. Aiming at solving the shortcoming of the present diagnostic method of intermittent fault, and according to the merit of support vector machines （ SVM） which can be trained with a small-sample, an SVM-based diagnostic model of 3 states that include OK state, intermittent state and faulty state is presented. With the features based on the reflection coefficients of an alarm rate （ AR ） model extracted from small vibration samples, these models are trained to diagnose intermittent faults. The experimental results show that this method can diagnose multiple intermittent faults accurately with small training samples and BIT false alarms are reduced.

快速凸包算法在发射车状态监控中的应用

针对当前复杂工业系统运行状态监控策略普遍存在误报警数目过多的问题,提出了一种基于快速凸包算法的发射车状态监控方法。该方法利用快速凸包算法从给定的正常历史数据中估计发射车的正常工作空间,对于新采集的发射车运行状态监控数据,如果由它们构成的工作点位于发射车的正常工作空间内,就认为发射车的状态是正常的,否则就是异常的。与基于静态阈值的方法相比,提出的方法降低了误报警数目;相较于基于隐变量的方法,提出的方法具有良好的可解释性。通过数值仿真技术,分别分析了一个2维和3维案例来评估提出的方法的性能表现。仿真结果表明,提出的方法物理意义明确,产生的误报警数目少。

基于多核PC的软件雷达信号积累和恒虚警处理研究

积累的本质是提高信噪比,以利于信号检测;恒虚警率处理的本质是在保证一定检测概率的基础上,用一个自适应门限代替固定门限,以保持虚警率的恒定,同时减小后端处理的压力,二者都是雷达信号处理系统的重要组成部分。在软件化雷达思想的指导下,针对雷达视频信号的积累和恒虚警率处理的实时性问题进行了研究。首先给出了积累和单元平均类恒虚警率处理的仿真模型,然后对积累和单元平均类恒虚警率处理的运算复杂度进行了详细的分析,接着采用集成性能原件(IPP)算法库和常规计算方法对积累和单元平均类恒虚警处理的运算时间进行仿真研究,最后通过实际软件实现,表明了在一定条件下,软件实现积累和恒虚警率可以满足实时性要求。