主要为数据分析和统计，解耦，恢复原信号，使用高阶累积量对MPSK信号进行调制识别，与理论分析结果相比，BP神经网络用于函数拟合与模式识别，使用起来非常方便，是机器学习的例程。- Mainly for data analysis and statistics, Decoupling, restore the original signal, Using high-order cumulants of MPSK signal modulation recognition, Compared with the results of theoretical analysis, BP neural network function fitting and pattern recognition, Very convenient to use, Machine learning routines.