特征提取方法在多源混合信号分选中的应用

在复杂电磁环境中,脉冲的大量丢失(数据缺失)及信号参数空间的严重交叠,破坏了传统分选方法所利用的信号规律性,最终导致现有信号分选算法很难获得令人满意的分选效果甚至完全失效。在同时考虑多个关键特征指标(脉冲宽度、载波频率、到达时间)的基础上,设计多参数五层互耦的分选算法;特别是提出新的关键特征指标提取方法,研究数据内部蕴含的特征向量,建立相应的数学模型,最终应用于对各种通信信号的分选。通过数值结果可以看出,引入的五步分选法可以实现对严重交叠和部分数据缺失的信号的分选。

Application of character extraction to signal sorting of multi-source mixed signals

In complex electromagnetic environment, a lot of signals are missing and many parameters of signals have severe superposition. The orderliness of signals are destroyed, making the existing sorting methods unsatisfactory and even disabled completely. This paper studies the important signal parameters(PW，RF and TOA) and introduces the five-step sorting methods. It is important that using the character extraction methods, the character vectors can be mined and the corresponding models of mathematics can be made. Finally, we can achieve all the kinds of signal sorting. The results of simulations show that the five-step sorting method can be used to sort the mixed signals with losing data and severe superposition.