In this paper,a method for spoofing detection based on the variation of the signal’s carrier-to-noise ratio(CNR)is proposed.This method leverages the directionality of the antenna to induce varying gain changes in th...In this paper,a method for spoofing detection based on the variation of the signal’s carrier-to-noise ratio(CNR)is proposed.This method leverages the directionality of the antenna to induce varying gain changes in the signals across different incident directions,resulting in distinct CNR variations for each signal.A model is developed to calculate the variation value of the signal CNR based on the antenna gain pattern.This model enables the differentiation of the variation values of the CNR for authentic satellite signals and spoofing signals,thereby facilitating spoofing detection.The proposed method is capable of detecting spoofing signals with power and CNR similar to those of authentic satellite signals.The accuracy of the signal CNR variation value calculation model and the effectiveness of the spoofing detection method are verified through a series of experiments.In addition,the proposed spoofing detection method works not only for a single spoofing source but also for distributed spoofing sources.展开更多
针对现有低功耗蓝牙(BLE)欺骗攻击检测技术准确率低的问题,提出了一种基于异常指纹的BLE欺骗攻击检测技术,将攻击者的射频指纹作为异常数据,把欺骗攻击检测建模为异常检测问题;设计了一种基于深度支持向量描述(Deep Support Vector Data...针对现有低功耗蓝牙(BLE)欺骗攻击检测技术准确率低的问题,提出了一种基于异常指纹的BLE欺骗攻击检测技术,将攻击者的射频指纹作为异常数据,把欺骗攻击检测建模为异常检测问题;设计了一种基于深度支持向量描述(Deep Support Vector Data Description,DeepSVDD)的异常指纹检测模型——RFFAD_DeepSVDD,并使用残差单元构建网络模型,有效缓解了机器学习异常检测算法非线性特征提取不足的问题。采用预训练自编码器获取最优初始化参数,极大增强了模型边界决策能力。在异常检测实验中,该模型准确率达到95.47%,相比基于机器学习的异常检测模型平均提升8.92%;在欺骗攻击检测实验中,该方法相比现有欺骗攻击检测技术在攻击节点运动与静止状态下均表现出更好的性能,能够准确检测并识别出中间人攻击、冒充攻击、重连接欺骗攻击3种欺骗攻击。展开更多
基金supported by the National Natural Science Foundation of China(62273195).
文摘In this paper,a method for spoofing detection based on the variation of the signal’s carrier-to-noise ratio(CNR)is proposed.This method leverages the directionality of the antenna to induce varying gain changes in the signals across different incident directions,resulting in distinct CNR variations for each signal.A model is developed to calculate the variation value of the signal CNR based on the antenna gain pattern.This model enables the differentiation of the variation values of the CNR for authentic satellite signals and spoofing signals,thereby facilitating spoofing detection.The proposed method is capable of detecting spoofing signals with power and CNR similar to those of authentic satellite signals.The accuracy of the signal CNR variation value calculation model and the effectiveness of the spoofing detection method are verified through a series of experiments.In addition,the proposed spoofing detection method works not only for a single spoofing source but also for distributed spoofing sources.
文摘针对现有低功耗蓝牙(BLE)欺骗攻击检测技术准确率低的问题,提出了一种基于异常指纹的BLE欺骗攻击检测技术,将攻击者的射频指纹作为异常数据,把欺骗攻击检测建模为异常检测问题;设计了一种基于深度支持向量描述(Deep Support Vector Data Description,DeepSVDD)的异常指纹检测模型——RFFAD_DeepSVDD,并使用残差单元构建网络模型,有效缓解了机器学习异常检测算法非线性特征提取不足的问题。采用预训练自编码器获取最优初始化参数,极大增强了模型边界决策能力。在异常检测实验中,该模型准确率达到95.47%,相比基于机器学习的异常检测模型平均提升8.92%;在欺骗攻击检测实验中,该方法相比现有欺骗攻击检测技术在攻击节点运动与静止状态下均表现出更好的性能,能够准确检测并识别出中间人攻击、冒充攻击、重连接欺骗攻击3种欺骗攻击。
