基于时频原子方法的雷达辐射源个体识别

A novel specific emitter identification method based on time-frequency atom approach

针对雷达辐射源个体识别问题,提出一种时频原子分解的特征提取方法。该方法针对主振放大式发射机,首先根据锁相环频率合成器相位噪声的功率谱,模拟含有特定相位噪声的辐射源信号。然后基于时频原子分解原理并采用膜算法将带有相位噪声的辐射源信号在正弦原子库中进行一次分解。最后利用分解后的原子提取出正弦原子能量特征和原子频率偏差特征,作为辐射源个体识别的分类特征。仿真实验验证了该方法的有效性。

A time-frequency atom decomposition based feature extraction method is proposed to implement the specific emitter identification. The method focuses on the master - oscillator power amplifier transmitter. First, the emitter signal with specific phase noise is simulated according to the power spectrum of the phase noise of phase-lock loop frequency synthesizer. Then, the emitter signal is decomposed once in the si- i nusoidal atom dictionary by the use of time-frequency atom decomposition and membrane algorithm. After that two features,called sinusoidal atom energy feature and carrier frequency bias feature, are extracted from the first decomposed atom as the features for specific emitter identification. The effectiveness of the proposed feature extraction approach is validated by the extensive experiments.