论文部分内容阅读
针对非线性盲源分离中非线性问题转化为线性问题,提出了一种基于AR模型的新方法。该方法在已知源信号的AR模型前提下,不但能够处理源信号的分离问题,还能够提取特定源信号,而后者是原来方法不具备的。从语音信号的非线性混合中提取源信号的仿真实验证实了该算法的有效性。
Aiming at the linear problem of nonlinear problem in nonlinear blind source separation, a new method based on AR model is proposed. The proposed method can not only handle the source signal separation problem but also extract the specific source signal under the premise of knowing the AR model of the source signal, which is not available in the original method. The simulation experiment of extracting the source signal from the non-linear mixture of speech signals proves the effectiveness of the algorithm.