论文部分内容阅读
对诱发电位(EP)信号中具有强脉冲过程的脑电图(EEG)噪声,可以用α稳定分布模型来描述。基于分数低阶矩对传统的Cohen类时频分布进行了改进,得到了新的分数低阶空间时频分布(FLO-STFM),据此提出了一种新的可在α稳定分布环境下工作的分数低阶空间时频欠定盲分离算法(FLO-TF-UBSS)。将该盲分离算法应用到EP信号的提取,仿真实验结果表明所提出的盲分离算法能较好地在EEG噪声环境下实现对EP信号的盲提取,相关系数以及盲提取效果都优于基于二阶的TF-UBSS算法。
Electroencephalogram (EEG) noise with strong impulsive processes in evoked potential (EP) signals can be described by the α-stable distribution model. Based on fractional lower moments, the traditional Cohen-like time-frequency distribution is improved and a new fractional lower-order space-time distribution (FLO-STFM) is obtained. Based on this, a new algorithm is proposed to work in a stable distributed environment Fractional Low-order Space Under-indefinite Blind Source Separation Algorithm (FLO-TF-UBSS). The blind separation algorithm is applied to the extraction of EP signals. The simulation results show that the proposed algorithm can blindly extract EP signals under EEG noise environment, and the correlation coefficients and blind extraction effects are better than those based on two Order TF-UBSS algorithm.