本文提出一個基於單通道腦電信號與機器學習的分類方法，用於達到準確偵測出測試者疲勞狀態的效果。由於以前的疲勞偵測系統大多數為以多通道腦電信號或是其他傳感方式為基礎，然而單通道腦電圖相較之下簡便許多且更具有成本效益，但是於實際情況下的準確度尚未評估。對睡眠與清醒狀態做二元分類可以相當容易的實現，然而昏昏欲睡的狀態則與前兩者有相當大的重疊，因此不容易將三種類別準確分類。為了解決這個問題，我們提出了一種階層半監督式的分類方法，以此達到有效地區分清醒、昏昏欲睡、以及睡眠三種狀態。而實驗結果顯示出我們的方法能夠準確地檢測困倦狀態。

This thesis proposes a single-channel EEG and machine-learning classification for drowsiness detection. Unlike previous systems that are based on multi-channel EEG or other sensing modalities, single-channel EEG is much simpler and more cost-effective, but its accuracy has not been assessed. Binary classification for asleep vs. awake state is well understood, but drowsy state overlaps both and has not been easy to classify accurately. To address this problem, we propose a two-level, semi-supervised classification method to effectively distinguish these three states. Experimental results show our approach to be able to accurately detect drowsiness.

Contents i
Acknowledgments v
1 Introduction 1
1.1 Motivation ............................................ 1
1.2 Drowsiness Sensing .................................... 2
1.3 Drowsiness Detection .................................. 2
1.4 Contributions ......................................... 2
1.5 Thesis Organization ................................... 3
2 Related Work 4
2.1 Background ............................................ 4
2.2 Drowsiness Detection .................................. 5
2.2.1 Physiological Measures-based Methods ................ 6
2.2.2 Behavioral Measures-based Methods ................... 7
3 System Overview 9
3.1 System Architecture ................................... 9
3.2 Node Subsystem ........................................ 9
3.2.1 MindWave Mobile ..................................... 10
3.3 Host Subsystem ........................................ 11
3.3.1 Data Collection ..................................... 11
3.3.2 Data Preprocessing and Classification ............... 13
4 Drowsiness Detection 15
4.1 Data Preprocessing .................................... 15
4.1.1 Extrema Filtering ................................... 16
4.1.2 Data Smoothing ...................................... 16
4.2 Feature Extraction .................................... 17
4.2.1 Time-Series Processing .............................. 17
4.2.2 Dimensionality Reduction ............................ 18
4.3 Classification ........................................ 19
4.3.1 Background .......................................... 20
4.3.2 Two-level classification ............................ 23
5 Evaluation 26
5.1 Experimental Setup .................................... 26
5.1.1 Experimental Environment ............................ 26
5.1.2 Experimental Data ................................... 26
5.2 Experimental Results .................................. 27
5.2.1 Performance with Different Feature Sets ............. 27
5.2.2 Performance with Different Machine-learning Models .. 29
5.2.3 Comparison between Time-series and Non-Time-series .. 32
5.2.4 Discussion .......................................... 34
6 Conclusions and Future Work 35
6.1 Conclusions ........................................... 35
6.2 Future Work ........................................... 35
6.2.1 Hardware ............................................ 36
6.2.2 Software ............................................ 36
6.2.3 Deep Learning ....................................... 36

[1] A. Cˇ olic´, O. Marques, and B. Furht, “Driver drowsiness detection and measurement methods,”
in Driver Drowsiness Detection, pp. 7–18, Springer, 2014.
[2] A. Sahayadhas, K. Sundaraj, and M. Murugappan, “Detecting driver drowsiness based on sensors:
a review,” Sensors, vol. 12, no. 12, pp. 16937–16953, 2012.
[3] T. Suprihadi, K. Karyono, et al., “Drowtion: Driver drowsiness detection software using mindwave,”
in Industrial Automation, Information and Communications Technology (IAICT), 2014
International Conference on, pp. 141–144, IEEE, 2014.
[4] D. Martinez-Maradiaga and G. Meixner, “Morpheus alert: A smartphone application for preventing
microsleeping with a brain-computer-interface,” in Systems and Informatics (ICSAI),
2017 4th International Conference on, pp. 137–142, IEEE, 2017.
[5] N. Gupta, D. Najeeb, V. Gabrielian, and A. Nahapetian, “Mobile ecg-based drowsiness detection,”
in Consumer Communications & Networking Conference (CCNC), 2017 14th IEEE
Annual, pp. 29–32, IEEE, 2017.
[6] K. T. Chui, K. F. Tsang, H. R. Chi, B. W. K. Ling, and C. K. Wu, “An accurate ecg-based transportation
safety drowsiness detection scheme,” IEEE Transactions on Industrial Informatics,
vol. 12, no. 4, pp. 1438–1452, 2016.
[7] K. Kurosawa, N. Takezawa, T. Sano, S. Miyamoto, M. Yasushi, and H. Hashimoto, “Drowsiness
prediction system for vehicle using capacity coupled electrode type non-invasive ecg measurement,”
in System Integration (SII), 2017 IEEE/SICE International Symposium on, pp. 306–311,
IEEE, 2017.
[8] T. C. Chieh, M. M. Mustafa, A. Hussain, S. F. Hendi, and B. Y. Majlis, “Development of vehicle
driver drowsiness detection system using electrooculogram (eog),” in Computers, Communications,
& Signal Processing with Special Track on Biomedical Engineering, 2005. CCSP 2005.
1st International Conference on, pp. 165–168, IEEE, 2005.
[9] X. Zhu, W.-L. Zheng, B.-L. Lu, X. Chen, S. Chen, and C. Wang, “Eog-based drowsiness detection
using convolutional neural networks.,” in IJCNN, pp. 128–134, 2014.
[10] J. Wongphanngam and S. Pumrin, “Fatigue warning system for driver nodding off using depth
image from kinect,” in Electrical Engineering/Electronics, Computer, Telecommunications and
Information Technology (ECTI-CON), 2016 13th International Conference on, pp. 1–6, IEEE,
2016.
[11] B. Akrout andW. Mahdi, “Yawning detection by the analysis of variational descriptor for monitoring
driver drowsiness,” in Image Processing, Applications and Systems (IPAS), 2016 International,
pp. 1–5, IEEE, 2016.
[12] L.-C. Shi, H. Yu, and B.-L. Lu, “Semi-supervised clustering for vigilance analysis based on
eeg,” in Neural Networks, 2007. IJCNN 2007. International Joint Conference on, pp. 1518–
1523, IEEE, 2007.
[13] Y. Choi, J. Park, and D. Shin, “A semi-supervised inattention detection method using biological
signal,” Annals of Operations Research, vol. 258, no. 1, pp. 59–78, 2017.