基于自相关函数三维熵的雷达信号分选

Radar Signal Sorting Based on Three Dimensional Entropy Features of Autocorrelation Function

针对当前雷达辐射源信号分选方法存在准确率不高和对噪声敏感的问题,提出一种新的分选方法,对接收到的信号首先分析其自相关函数,然后提取该自相关函数的香农熵、指数熵和范数熵,并将该三维熵特征作为分选参数,最后基于KFCM算法实现分选。由于不同信号的自相关函数区别大且对噪声不敏感,因此提取的三维熵特征具有较好的分离性和稳定性。仿真结果表明,虽然雷达辐射源信号的周期性和自相关函数的最大延迟数对分选性能均存在一定的影响,但该方法依然具有较好的适应性和有效性。

The current sorting methods for radar emitter signals have not high accuracy and they are sensitive to the signal-noise-rate（SNR）as well.To solve these problems,a novel sorting method is proposed.Firstly,the autocorrelation function of the received signal is analyzed.Then Shannon entropy,exponent entropy,and norm entropy of its autocorrelation function are extracted and used as the sorting parameters.Finally,the sorting is accomplished by KFCM algorithm.The autocorrelation functions of different signals are different and not sensitive to SNR,so the extracted three dimensional entropy features have good separability and stability.The simulation results show that the periodicity of radar emitter signal and maximum delay number of the autocorrelation function have some effects on the sorting performance,but this method has good adaptability and effectiveness.