【摘要】 本文根据准平稳信号平稳段内信号功率频谱密度保持不变的特性,利用Kohonen自组织神经网络,提取准平稳信号的段数、段类、段序三个特征,从而有效地识别不同的准平稳信号。最后经过计算机模拟实验证明了该方法的有效性。更多还原

【Abstract】 According to the invariability of power spectrum of stationary signal segment, Kohonen’s self-organizing neural network is applied to find three features in quasi-stationary signal: the number of segment, the kind of segment and the order of segment. The pattern recognition of quasi-stationary signal is accomplished efficiently with these features.更多还原