改进的基于特征函数的调制识别算法

原有的基于特征函数的调制识别算法,是将基带信号的同相分量和正交分量之和代入特征函数进行运算,计算过程丢失了部分有用信息,仅能保证特征函数的相位跟高斯白噪声无关。提出一种新方法,将基带信号以复数形式代入特征函数进行运算,不仅没有丢失信号的有用信息,并且从理论上证明计算结果跟高斯白噪声无关,继而可以使用特征函数的幅度信息对信号的调制方式进行识别。仿真证明,该算法比原有方法在识别性能上有大幅提高。

Modified modulation classification method based on characteristic function

In existing modulation classification algorithm based on characteristic function, a real number which is usually the sum of the signal in-phase component and quadrature component is taken into the formula. In this process some signal information is lost, though the phase of characteristic function is insensitive to the white Gaussian noise. A new algorithm is proposed for calculating the characteristic function, which takes the complex baseband signal into the formula directly. This process does not lose any signal information. It is proved that the proposed characteristic function is insensitive to the white Gaussian noise, so the amplitude of the characteristic function can be used for modulation classification. Simulation result shows that the recognition performance is the improved remarkably.