基于谱相关函数截面特征的调制方式识别

Modulation Recognition based on Profile Features of Spectral Correlation Function

谱相关函数在双频平面上全面反映了循环平稳信号的二阶周期性。调制信号是循环平稳信号,不同调制信号的谱相关函数具有不同的分布特性,且对平稳噪声、干扰不敏感,可用于低信噪比下的调制方式识别。采用时域平滑FFT积累算法计算出调制信号的谱相关函数,对结果进行了对比分析,提取了多种调制信号的谱相关函数截面特征,设计了AM、MSK、PAM、2ASK、2FSK、4FSK、QPSK七种调制信号的树形识别流程。仿真结果表明在低信噪比下,七种调制方式具有90%以上的识别率。

Spectral correlation function（ SCF） fully reflects second-order periodicity of cyclostationary signals in the dual frequency plane. Modulation signals are cyclostationary signals and SCF is highly distinct for them. Because it is insensitive to stationary noise and interference,SCF can be used to classify modulation signals even under low SNR. SCF of several modulated signals are calculated using the FFT accumulation algorithm. By analyzing the results,features in different sections are extracted. Then a simple recognition process is designed for seven kinds of modulation signals,including AM,MSK,PAM,2ASK,2FSK,4FSK,QPSK. Simulation results show that recognition rate is better than 90% under low SNR.