中文说明:BP神经网络用于函数拟合与模式识别,包含特征值与特征向量的提取、训练样本以及最后的识别,使用高阶累积量对MPSK信号进行调制识别,分析了该信号的时域、频域、倒谱,循环谱等,模式识别中的bayes判别分析算法,基于chebyshev的水声信号分析。
English Description:
BP neural network for function approximation and pattern recognition, including eigenvalue and eigenvector extraction, training samples and the final recognition, using high order cumulants for modulation recognition of MPSK signals, analyzes the the signal in time domain and frequency domain, cepstrum, cyclic spectrum, pattern recognition in Bayes discriminant analysis algorithm, based on Chebyshev underwater acoustic signal analysis.