基于多分辨率动态模态分解的电磁信号时频-能量分析

针对电磁信号时-频-空间-能量分析不兼容的问题, 提出一种基于分解时空域电磁辐射信号的数据驱动方法来分析复杂电磁环境。首先, 使用多分辨率动态模态分解方法将复杂电磁信号以不同的分辨率分离为不同时空尺度的模态分量。然后，利用时频分布对各模态瞬时频率和能量特征进行采集。这种动态的分解分析方式能快速侦测电路和电子系统的多方面电磁环境状态, 具有不需要先验数据训练, 不受时频分辨率限制, 可以实时分析和降低数据维数等优点，且提取的动态模式及其频谱分别在空间域和时域具有明确的物理意义。仿真实验显示, 重组信号和原信号的平均误差在-18 dB以下, 证明了所提方法的可行性和优势。

Time-frequency-energy analysis of electromagnetic signals based on multi-resolution dynamic modal decomposition

Aiming at the incompatibility of time-frequency-space-energy analysis of electromagnetic signals, a data-driven method based on decomposing electromagnetic radiation signals in space-time domain is proposed to analyze complex electromagnetic environment. Firstly, the multi-resolution dynamic mode decomposition method is used to separate the complex electromagnetic signal into modal components of different temporal and spatial scales at different resolutions. Then, the instantaneous frequency and energy characteristics of each mode are collected by using time-frequency distribution. This dynamic decomposition analysis method can quickly detect various electromagnetic environment states of circuits and electronic systems. It does not need a priori data training, is not limited by time-frequency resolution, and can analyze and reduce the data dimension in real time. The extracted dynamic mode and its spectrum have clear physical significance in space domain and time domain respectively. Simulation results show that the average error between the reconstructed signal and the original signal is less than -18 dB, which proves the feasibility and advantages of the proposed method.