基于高阶累积量和SVM的OFDM调制制式识别  

Modulation Recognition of OFDM Based on Higher Order Cumulants and SVM

作　　者：王建勇[1] 王海[2] 

机构地区：[1]华中农业大学理学院,武汉430070 [2]武汉数字工程研究所,武汉430074

出　　处：《计算机与数字工程》2008年第6期41-44,共4页Computer & Digital Engineering

基　　金：华中农业大学创新基金(编号:52204-04078)资助

摘　　要：针对Rayleigh衰落信道条件下OFDM信号与单载波信号调制识别问题,提出一种基于高阶累积量和支撑矢量机的信号识别算法。在接收端计算接收信号的高阶累积量作为提取的特征值,构造分类特征向量,支撑矢量机分类器将分类特征向量非线性地映射到高维特征空间中,并建立一个最优超平面来实现信号调制方式的分类。该方法不需要信号的载波频率、波特率等先验信息,直接在中频对信号进行处理。仿真结果表明,该算法对Rayleigh衰落和高斯噪声均不敏感,对信噪比的变化有很强的适应能力。且该算法识别概率高,性能稳定,复杂度低,适合实时处理。This paper presents a new method to discriminate OFDM from single - carder modulations based on higher order cumulants and support vector machines（SVM） over a Rayleigh fading channel. The classification feature vector is extracted from higher order cumulants of the received signal. SVM maps input vectors nonlinearly into a high dimensional feature space and con- structs the optimum separating hyperplane in the space to realize signal recognition. The algorithm directly processes the signals on intermediate frequency without the prior information about the carrier frequency and the symbol rate of received signal. Simulation results demonstrate that this method is robust to the Rayleigh fading, Guass noise and variety of SNR. Moreover, the high correct classification rate, robustness and low complexity are proved by computer simulation.

关 键 词：正交频分复用 支撑矢量机 高阶累积量 调制识别 

分 类 号：TP393.04[自动化与计算机技术—计算机应用技术]