复杂欺骗环境下GNSS欺骗信号特征提取算法设计与分析

Design and Analysis of GNSS Spoofing Signal Feature Extraction Algorithms under Complex Spoofing Environments

摘要:

针对GNSS信号在受欺骗攻击时易受到欺骗信号的影响的问题,研究GNSS欺骗信号频域特征提取算法设计和统计特性。根据GNSS信号频谱结构和高阶统计特性的变化特征,给出了频域参数、频域矩峰度系数、单频能量聚集度、频域矩偏度系数、平均频谱平坦系数五种频域特征,给出了相应的数学模型和计算方法,通过仿真信号,对正常与欺骗条件下各特征的统计分布情况进行了对比,通过区间统计方法来反映特征在不同样本序号和实验条件下的波动程度,并以Cohen’s d指标进行定量分析。得出不同频域特征在正常和欺骗条件下各特征间具有一定的统计差异,部分特征具有较强的区分能力,能够反映欺骗信号对GNSS频谱统计特性的影响。研究结果为复杂欺骗环境下GNSS欺骗检测算法特征的选择和设计提供了理论依据。

Abstract:

To address the vulnerability of GNSS signals to deception attacks, this study investigates the design and statistical characteristics of frequency domain feature extraction algorithms for GNSS deception signals. Based on the variation characteristics of the GNSS signal spectral structure and higher-order statistical properties, five frequency domain features are identified: frequency domain parameters, frequency domain moment kurtosis coefficient, single-frequency energy concentration, frequency domain moment skewness coefficient, and average spectral flatness coefficient. Corresponding mathematical models and calculation methods are provided. Through simulated signals, the statistical distribution of each feature under normal and deception conditions is compared. Interval statistics are used to reflect the fluctuation degree of features under different sample indices and experimental conditions, and Cohen"s d" index is used for quantitative analysis. The results show that different frequency domain features exhibit certain sta-tistical differences between normal and deception conditions, with some features possessing strong discriminative power and re-flecting the influence of deception signals on the statistical characteristics of the GNSS spectrum. The research results provide a theoretical basis for the selection and design of features in GNSS deception detection algorithms under complex deception envi-ronments.