采用高阶累计量的时频混叠信号调制识别研究

随着当今无线通信技术的发展，频谱资源的划分日益紧张，在同一信道里出现两个或两个以上时频混叠信号的现象越来越普遍。文章针对共信道时频混叠信号调制识别的问题，提出一种基于高阶累计量的算法。该算法分别对混叠信号中各个信号以其参数进行载波同步与定时同步后由高阶累积量计算混叠信号的累积量处理得到混叠信号的特征参数，而后由SVM（支撑向量机）分类器进行多分类识别。经仿真验证，该算法能够实现BPSK，QPSK，8PSK，8QAM，16QAM信号混合后的识别。在对比已有文献的基础上，严格证明了经过载波同步和定时同步后混叠信号处理等效为单信号的结论，简化了特征参数的选取，同时采用SVM优化了分类性能并简化了分类流程。

Method of Modulation Recognition of Time-Frequency Overlapped Signals Based on High-Order Cumulants

With the current rapid development and wide application of radio communication technol- ogy, signal environment is getting more and more complex, it＇ s more common for two or more time- frequency overlapped signals to appear in single channel. To solve the modulation recognition prob- lem of time-frequency overlapped signals, this paper presents a modulation classification method based on high-order cumulants. This algorithm makes carrier synchronization and timing synchroni- zation through its single signal, calculates the co-channel signal high-order cumulants. The charae- teristic parameter of high-order cumulants of the co-channel signals is used for training data and tes- ting data based on SVM （Support Vector Machine） to he classifed. The simulation proves that the method is able to identify BPSK, QPSK, 8PSK, 8QAM, 16QAM after mixing. This paper, com- pared to the existing literature, strictly proves the theory that after carrier synchronization and timing synchronization, the characteristic of high-order cumulants of the time-frequency overlapped signals is equal to that of the single signal. Through this algorithm the parameters are simplified, and SVM optimizes classification performance and simplifies the classification process with a high perform- ance.