使用高阶累积量对MPSK信号进行调制识别，实现了对10个数字音的识别程序包括AHP，因子分析，回归分析，聚类分析，可以得到很精确的幅值、频率、相位估计，一种流形学习算法（很好用），实现典型相关分析，结合PCA的尺度不变特征变换(SIFT)算法。- Using high-order cumulants of MPSK signal modulation recognition, Realization of 10 digital audio recognition program Including AHP, factor analysis, regression analysis, cluster analysis, You can get a very accurate amplitude, frequency, phase estimation, A fluid manifold learning algorithm (good use), Achieve canonical correlation analysis, Combined with PCA scale invariant feature transform (SIFT) algorithm.