Abstract: This paper proposes a method to identify the complex radar signals，such as phase-shift keying signal，frequency-shift keying signal and hybrid modulation signal.Based on the combination of AR model power spectrum estimation and modulus maxima of wavelet transform，the complex modulated signals are converted into phase-shift keying signals.Then the frequency characteristics of the signals are extracted by high-order spectrum analysis.Simulation results show that the proposed method can identify up to six kinds of complex radar signals，and most signals can achieve more than 93% recognition accuracy when SNR is above 0 dB.

Key words: hybrid modulation signal, AR model power spectrum estimation, modulus maxima, high-order spectrum analysis

摘要： 提出一种识别包括相位编码、频率编码，以及脉内混合调制等复杂雷达信号的方法。该方法结合AR模型功率谱估计分析方法和小波变换模极大值方法，将复杂混合调制信号转化成相位编码信号，然后基于高次方频谱分析的方法提取信号的频率特征，首次实现了对6种复杂雷达信号的分类。实验表明，在信噪比为0 dB时，大多数信号的识别率可以达到93%以上。

关键词: 脉内混合调制雷达信号, AR模型功率谱分析, 模极大值, 高次方频谱分析