基于滑动窗独立分量分析的在线包络检测新方法及其在脑-机接口中的应用  

The Online Envelope Detection Based on Sliding Window ICA and Its Application to Brain-Computer Interface

作　　者：吴小培[1] 宋俊可[1] 郭晓静[1] 巩笑晓[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,合肥230039

出　　处：《生物物理学报》2012年第11期896-909,共14页Acta Biophysica Sinica

基　　金：国家自然科学基金项目(60771033;61271352);博士点基金项目(200803570002)~~

摘　　要：独立分量分析(Independent component analysis,ICA)是一种有效的空域滤波方法,在脑-机接口领域具有很高的研究价值和应用潜力。论文围绕基于滑动窗的自适应ICA算法及其在脑-机接口中的应用开展研究,提出了一种基于ICA的信号包络在线检测新方法,并将该方法应用于运动想象诱发的mu节律包络检测,建立了相应的左右手运动想象分类识别算法。在此基础上,进行了离线和在线运动想象脑-机接口实验,取得了较好的识别效果。Independent component analysis （ICA） is an effective spatial filtering method and has promising potentials in the field of brain computer interface （BCI）. This paper focuses on the adaptive ICA algorithm based on sliding window, and a novel approach of ICA-based envelope detection is proposed, which is applied to the envelope detection of mu rhythm induced by motor imagination and the corresponding classification algorithm is developed. With this algorithm, the offline and online experiments of BCI of movement imagery are carried out, and the good recognition results are obtained.

关 键 词：独立分量分析 包络检测 脑-机接口 滑动窗 运动想象 

分 类 号：R318[医药卫生—生物医学工程]