本研究開發利用模糊理論於穩態視覺誘發(Steady-State Visual Evoked Potential, SSVEP)之大腦人機界面(Brain Computer Interface, BCI)判斷。受測者注視閃光頻率為32Hz且不同相位的4個閃光，使大腦誘發出相對應32Hz的穩態視覺誘發電位。使用數位22Hz~42Hz帶通濾波器濾腦波訊號(electroencephalography, EEG)，並利用疊加平均的方法處理訊號，為判斷做前置處理。訊號處理的末端利用模糊理論來對每個不同受測者的腦波訊號做最佳化判斷，並把所有對腦波訊號的訊號處理及決策以微處理器硬體實現，最後把判斷的結果用無線藍芽模組傳輸到電腦用LabVIEW收結果，並藉以控制人形機器人，實現受測者所要下達給機器人的指令。實驗結果發現受測者用模糊理論判斷方法比一般選項判斷方法，其準確率提高1~12%，平均準確率為93.38 ± 5.74%，平均的ITR為60.26 ± 23.71 bits/min，從結果可以明顯看出模糊判斷方法可以改善一般判斷方法的不足，進而提高準確率。The thesis applied Fuzzy Theory to the judgment of steady-state visual evoked potential (SSVEP)-based brain computer interface (BCI).User gazed at flash channels(FCs) that encoded with different phases in order to induce the corresponding SSVEP , so that the gazed FC can be recognized and the command mapping to the gazed FC can be sent out to achieve control purposes. In the thesis, the frequency of FC is 32 Hz, and there are four FCs with different phases 0ﾟ, 90ﾟ, 180ﾟ and 270ﾟ. The SSVEP responses were processed by 20–36 Hz filter and epoch-average.Using Fuzzy Theory to optimize the judgment of BCI system can reduce the occurrence of error judgments. We use a micro-processor to do all the signal process about electroencephalography (EEG), and transmit the result to PC with Bluetooth. PC will sent out the control direction to the robot, and the robot do the actions that user wants.The experiment results show that apply Fuzzy Theory to the judgment of BCI system can increase more 1~12% accuracy than normal judgment theory. Some subjects’ accuracy can even reach 100% by using Fuzzy Theory.
