VHF/UHF常用通信信号调制识别算法

摘要

针对甚高频、超高频频段内通信信号的调制方式种类繁多,对其进行类间识别与类内识别难度大以及在低信噪比时性能不理想的问题,提出了一种基于特征提取的调制识别方法.选取了一组稳健性强的特征参数,综合利用接收信号的包络高阶特征、非线性变换后的谱特征、瞬时频率聚类特征和改进的平坦度指数,将二叉树与支持向量机相结合构造分类器,对该频段内常见的多种信号完成了调制识别.算法无需任何先验知识,不受调制方式参数变化的影响具有较强的抗噪声能力,仿真实验表明了算法的有效性和正确性.%The modulation mode category of communication signal is various in the very high frequency and uhrahigh frequency.The difficulty of inter-class recognition and intra-class recognition is huge and the performance during low signal to noise ratio is unsatisfactory.In this paper,we proposed a modulation recognition method based on the feature extraction aiming at the problem mentioned above.Firstly,we selected a group of feature parameters with strong robustness.Then,we used the higher order envelope characteristic of received signal,spectrum characteristic after nonlinear transformation,clustering characteristic of instantaneous frequency,and the modified flatness index.Finally,we integrated the binary tree with the support vector machine to build the classifier and completed the modulation recognition on various common signals in the frequency band.The algorithm dispenses with any prior knowledge and is unaffected with parameter variation of modulation mode.It has better noise resisting ability.The simulation results indicate the effectiveness and validity of algorithm.