稳定分布噪声环境下的EEG-MUSIC源定位

EEG-MUSIC source localization in α stable noise environment

EEG作为一种无损的新医学技术近几年已经被广泛地研究,EEG-MUSIC是一种基于EEG的脑源病灶空间定位算法,它只能在高斯噪声下工作。特殊情况下的脑电传感器阵列数据中可能伴有一种强脉冲噪声,这种噪声可以用α稳定分布描述。稳定分布噪声环境下,传统的EEG-RAP-MUSIC失效。为此用FLOM矩阵代替自相关矩阵,提出了一种适合稳定分布噪声环境的脑源定位新方法FLOM-EEG-RAP-MUSIC。计算机仿真表明,所提出的方法能较好地在稳定分布噪声环境下工作,具有一定的鲁棒性。

EEG（Electroencephalography） which is a new medical technology has been generally researched recently, EEG-MUSIC （Multiple Signal Classification） is a spatial source localization method based on EEG, but it can only work in Gauss noises environment. Because noises are pulsing in EEG array processing data, they can be described by a stable distribution. The per- formance of EEG-RAP （Recursively Applied and Projected）-MUSIC degenerates in a stable distribution environment. A new source localization method which can work in a stable distribution environment is proposed because correlation matrix is re- placed by improved Fractional Low Order Moment（FLOM）matrix in the method. Simulations indicate that the method is robust.