利用相对小波能量和概率网络的脑-机接口

Brain-computer interface using RWE and PNN

脑-机接口是一种全新的人机接口方式,在人脑与计算机或其他电子设备之间建立的直接的交流和控制通道。特征提取和分类是脑-机接口的关键。脑电信号经过预处理后,利用脑电信号的相对小波能量作为特征,采用主分量分析进行降维,然后利用概率神经网络对两类不同的意识任务(想象小手指运动和舌头运动)进行分类。离线分析结果表明,该方法在分类准确率上有很大的提高,从而为脑-机接口系统的特征提取和分类提供了新思路。

The brain-computer interface（BCI） is a novel kind of human computer interface and provides a direct communication and control channel for sending messages and instructions from brain to external computers or other electronic devices.Feature selection and classification are the key points in BCI research.Relative wavelet energy（RWE） of EEG signal after preprocessing is used as feature.Then the feature vector is reduced using principal components analysis （PCA）.Probabilistic neural network （PNN） is used to classify two different mental tasks （imaged litter finger and tongue movement ）.From the off-line result of the experiment,this algorithm has got significant improvement on classification accuracy,which can be expected to provide a new way for the feature selection and classification in BCI design.