以經驗模態分解法分析穩態視覺誘發電位之大腦人機界面; Implementation of a steady-state visual evoked potential based brain computer interface using Empirical Mode Decomposition

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Electrical Engineering > Electronic Thesis & Dissertation > Item 987654321/10435

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/10435

Title:	以經驗模態分解法分析穩態視覺誘發電位之大腦人機界面;Implementation of a steady-state visual evoked potential based brain computer interface using Empirical Mode Decomposition
Authors:	李昆興;Kung -Shing Li
Contributors:	電機工程研究所
Keywords:	大腦人機界面;視覺誘發電位;Quadrature Detection;Zero-Crossing;經驗模態分解法;Brain computer interface (BCI);Steady-state visual evoked potential (SSVEP);Empirical Mode Decomposition (EMD);Zero-Crossing(ZC);Quadrature Detection(QD)
Date:	2009-06-28
Issue Date:	2009-09-22 12:16:55 (UTC+8)
Publisher:	國立中央大學圖書館
Abstract:	近年來，視覺誘發電位(steady state visual evoked potential, SSVEP)為基礎之大腦人機界面已被廣泛使用，利用不同頻率進行編碼以表示各種控制指令，透過非侵入式腦波訊號(Electroencephalogram, EEG)的擷取及辨識，達到輔助嚴重神經肌肉損傷患者，成為與外界溝通的橋樑。SSVEP-BCI具有高傳輸率與少訓練時數等特點，但是腦波訊號屬於非線性(nonlinear )、非穩態(non-stationary)且易受雜訊干擾的隨機程序，因此本研究利用經驗模態分解法(Empirical Mode Decomposition, EMD)自適性(adaptive)之特性移除訊號中非週期、不連續與非事件相關等雜訊，建立快速辨識之大腦人機界面。在本系統中使用6組中頻率(≧30Hz)閃爍LED燈做為視覺刺激材料以誘發SSVEP，腦波訊號先由EMD根據瞬時頻率的概念將訊號拆解成不同時間尺度之內建模式函數(intrinsic mode function, IMF)，並以Zero-Crossing(ZC)估測其頻率挑選出符合刺激頻率之成份以重建SSVEP訊號，重建之腦波訊號從時域上利用Quadrature detection(QD)直接地解調出使用者所注視頻率，再經由電腦轉譯以表達使用者的意識進行與外界的溝通，本研究將BCI系統建構於LabVIEW平台具體實現即時且連續的控制目的。本BCI系統目前應用於模擬滑鼠移動的6個動作選項，在offline的穩定度測試中，5位受測者對於六個選項(30~35 HZ)的平均資料傳送率(Information transfer rate, ITR)達到60 bits/min，而在滑鼠連續移動測試的online實驗中則達到平均30 bits/min。 In recent years, the brain computer interface (BCI) based on the steady-state visual evoked potential (SSVEP) has been widely used in many applications. By tagging of flickers with different frequencies, user’s gazed target can be recognized by analyzing the frequencies of evoked SSVEPs. Though SSVEP - based BCI has the advantages of high information transfer rate (ITR) and less training time, the extraction of SSVEP is sometimes not complete due to its characteristics of nonlinearity, non-stationary and noise susceptibility. Accordingly, this study adopts Empirical Mode Decomposition (EMD) to remove SSVEP-unrelated noise and tries to implement a rapidly – controlled SSVEP-based BCI system. In this study, six LEDs with high-flickering rates (30, 31, 32, 33, 34, and 35 Hz) were utilized as visual stimulators to induce subjects’ SSVEPs. EEG signals recorded at Oz channel were decomposed by EMD into a series of oscillation components, representing muliti-scale information of the signal, called intrinsic mode functions (IMFs). The frequency of each IMF, denoted as fzc, was examined by zero crossing (ZC), and only those IMFs, which had their fzc close to the flickering frequencies of the six LEDs (30.5Hz ? fzc ? 36.5Hz), were designated as SSVEP-related IMFs for reconstruction noise-suppressed SSVEPs. The proposed system has been used to control a cursor with six cursor functions. It has achieved high ITR (> 30 bits/min) in online controls.
Appears in Collections:	[Graduate Institute of Electrical Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Size	Format

社群 sharing

Loading...