摘要
脑-机接口是一种全新的信息交换与控制技术,其关键技术是要准确及时地识别出思维脑电的预期动作模式及其特征参数并将之转化为控制命令。本文通过分析和提取左、右手两种想象动作诱发的事件相关去同步(ERD)/同步(ERS)信号特征,并采用BP神经网络对特征量进行模式分类识别,以区分想象动作的意图,进而形成可用来控制轮椅、鼠标等外部机电设备的控制指令。研究结果表明该方法识别率高,速度快,可用于实时脑—机接口系统中。
Brain-computer interfaee（BCI） is a novel information interchange and control technology,which works mainly accurately and promptly identifying the expected action patterns and characteristic parameters of the mental EEG and converting them to control orders. The characteristics of event related desynchronization/synchronization（ERD）/（ERS） signals that are evoked by imaging left hand and right hand actions are analyzed and extracted and then the BP neural network is used to identify the characteristics by class to distinguish directions of imagined actions and produce command to control external devices like wheelchair,mouse etc. The research results show that BP neural network performs well in recognition rate and speed,thus it can be applied to real-lime BCI system.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2008年第5期676-679,共4页
Journal of Optoelectronics·Laser
基金
国家自然科学基金资助项目(60471028)
天津市教委基金资助项目(20051215)
关键词
脑-机接口
模式识别
BP神经网络
想象动作电位
事件相关去同步/同步
brain-computer interface
pattern recognition
BP neural networks
motor imaginary potentials
event related desynchronization/synchronization
