駕駛疲勞是車禍中常見的肇事原因，而生理訊號被認為是偵測駕駛疲勞最可靠的方法。本論文提出基於心率變異性的疲勞駕駛偵測系統，會著重在疲勞事件的標記及分類。在這篇論文中，我們使用精神運動警覺性任務 (PVT)及Karolinska睏睡度量表 (KSS)來測量駕駛員的狀態，使用四個參數來當作標記疲勞事件，分別是反應時間 (RT)、失誤 (lapse)次數、∆RT >250毫秒的次數及KSS。疲勞閾值的選擇使用Wilcoxon等級和檢定，將資料區分成最有顯著差異的兩組，並將對應到的值作為疲勞事件的閾值。
我們設計標記實驗中所對應心電圖資料的方法，相鄰十分鐘的行車模擬實驗會被標註成和最近的PVT狀態相同。最後，使用被標記的心率變異性特徵當作支持向量機的輸入，得出分類結果。另外，我們還使用了真實道路的實驗作為提出模型的驗證。
相較於其他文獻使用時間或臉部特徵當作疲勞參考，我們採用的依據分別為駕駛員主觀及客觀的結果。使用KSS參數的模型準確度為94.52 %，其他使用PVT相關參數的模型準確度分別為91.39 %、90.95 % 和90.31 %。此外，只使用PVT階段的KSS參數模型準確度為97.79 %。我們提出的標記參數反映了駕駛員真實的感受及反應，提出的模型和其他文獻相比，有更好的準確度也更適合應用在真實情況的駕駛疲勞偵測系統。然而，當數據以人為單位分割時，模型準確度僅能達到80%，顯示個人化差異仍存在於模型中。
大多數文獻缺少的額外道路實驗在本論文中被用來驗證提出的系統，結果顯示，使用KSS參數的模型比其他參數得到更高的準確度，分別為81.09 % 和86.51%。總結來說，本論文不僅有明確的疲勞標籤，也得到比大部分文獻還要好的結果。

Driving fatigue is one of the critical reasons in car accidents and physiological signals are considered the most reliable method in detecting driving fatigue. This thesis proposed a driver fatigue detection system based on heart rate variability (HRV) via support vector machine (SVM). We focused on how to label and classify the fatigue events. In this thesis, we used the psychomotor vigilance task (PVT) and Karolinska sleepiness scale (KSS) to measure the state of drivers. The parameters adopted to label fatigue events are reaction time (RT), the number of lapses, the number of $\Delta$ RT $>$ 250ms, and the KSS level, respectively. The thresholds of these parameters are obtained by the Wilcoxon rank sum test. Then we designed label criteria to label the corresponding electrocardiogram (ECG) data of our experiments. The adjacent 10 minutes driving data were considered as the same state as PVT stages. Finally, the 5-minute HRV features extracted from ECG signals were labeled as the input of SVM models to classify fatigue or not. Furthermore, we demonstrated the performance of our SVM model to predict the occurrence of fatigue using 5-minute HRV signals by the real road test with the same experimental process.

Compared to other works, we adopted fatigue references that related to subjective and objective measures of subjects instead of using the facial features or state over time. The results of SVM models with the KSS label were 94.52%. The results of SVM models with PVT-related parameters were 91.39%, 90.95%, and 90.31%, respectively. In addition, the accuracy of the SVM models with the KSS label using only PVT stages data was 97.79%. Even though the collection procedure of our data had to be designed specifically, our labeled parameters reflected the real subjective feeling and reactions from drivers. Our models presented with higher accuracy than other works and could be more suitable in realizing a real driving detection system. However, the personal difference inevitably existed in our models since the accuracy of our model can only reach 80% when data were split by the person.

The extra real road tests which most works lacked were also used as verification. The results turned out that our fatigue detection system with the KSS label performed the best among others. The mean accuracy was 81.09% and 86.51%. In conclusion, this thesis had not only clear fatigue labels but also better and reliable results than most works.

[1] A. W. MacLean, D. R. Davies, and K. Thiele, “The hazards and prevention of driving while sleepy,” Sleep Medicine Reviews, vol. 7, no. 6, pp. 507–521, 2003.

[2] “Car accidents from drowsy driving statistics,” 2020. [Online]. Available: https: //www.nhtsa.gov/risky-driving/drowsy-driving.

[3] “2019 traffic safety culture index (technical report).” 2019. [Online]. Available: https://aaafoundation.org/2019-traffic-safety-culture-index/.

[4] F. Wang, J. Lin, W. Wang, and H. Wang, “EEG-based mental fatigue assessment during driving by using sample entropy and rhythm energy,” IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems (CYBER), pp. 1906– 1911, 2015.

[5] E. Rogado, J. Garcia, R. Barea, L. Bergasa, and E. Lopez, “Driver fatigue detection system,” 2008 IEEE International Conference on Robotics and Biomimetics, pp. 1105– 1110, 2009.

[6] R. Buendia, F. Forcolin, J. Karlsson, B. A. Sj¨oqvist, A. Anund, and S. Candefjord, “Deriving heart rate variability indices from cardiac monitoring—an indicator of driver sleepiness,” Traffic Injury Prevention, vol. 20, no. 3, pp. 249–254, 2019.

[7] G. Attarodi, S. Matla Nikooei, N. Jafarnia Dabanloo, P. Pourmasoumi, and A. Tareh, “Detection of driver’s drowsiness using new features extracted from HRV signal,” Computing in Cardiology Conference (CinC), vol. 45, pp. 1–4, 2018.

[8] J. Vicente, P. Laguna, A. Bartra, and R. Bail´on, “Detection of driver’s drowsiness by means of HRV analysis,” Computing in Cardiology, pp. 89–92, 2011.

[9] Y.-L. Chang, Y.-C. Feng, and O. T.-C. Chen, “Real-time physiological and facial monitoring for safe driving,” 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 4849–4852, 2016.

[10] M. Awais, N. Badruddin, and M. Drieberg, “A hybrid approach to detect driver drowsiness utilizing physiological signals to improve system performance and wearability,” Sensors, vol. 17, no. 9, 2017.

[11] K. Fujiwara, E. Abe, K. Kamata, C. Nakayama, Y. Suzuki, T. Yamakawa, T. Hiraoka, M. Kano, Y. Sumi, F. Masuda, M. Matsuo, and H. Kadotani, “Heart rate variability based driver drowsiness detection and its validation with EEG,” IEEE Transactions on Biomedical Engineering, vol. 66, no. 6, pp. 1769–1778, 2019.

[12] Tunc¸er, L. G¨uvenc¸, F. Cos¸kun, and E. Karslıgil, “Vision based lane keeping assistance control triggered by a driver inattention monitor,” IEEE International Conference on Systems, Man and Cybernetics, pp. 289–297, 2010.

[13] P. Thiffault and J. Bergeron, “Monotony of road environment and driver fatigue: a simulator study,” Accident Analysis Prevention, vol. 35, no. 3, pp. 381– 391, 2003.

[14] R. Sayed and A. Eskandarian, “Unobtrusive drowsiness detection by neural network learning of driver steering,” Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, vol. 215, no. 9, pp. 969–975, 2001.

[15] J. Verster and T. Roth, “Standard operation procedures for conducting the on-the-road driving test, and measurement of the standard deviation of lateral position (SDLP),” International journal of general medicine, vol. 4, pp. 359–71, May. 2011.

[16] H. Han, H. Jang, and S. W. Yoon, “Driver head posture monitoring using MEMS magnetometer and neural network for long-distance driving fatigue analysis,” IEEE Sensors, pp. 1–4, 2019.

[17] M. Ali, S. Abdullah, C. S. Raizal, K. F. Rohith, and V. G. Menon, “A novel and efficient real time driver fatigue and yawn detection-alert system,” 3rd International Conference on Trends in Electronics and Informatics (ICOEI), pp. 687–691, 2019.

[18] H. Singh, J. S. Bhatia, and J. Kaur, “Eye tracking based driver fatigue monitoring and warning system,” India International Conference on Power Electronics, pp. 1–6, 2011.

[19] A. Sahayadhas, K. Sundaraj, and M. Murugappan, “Detecting driver drowsiness based on sensors: A review,” Sensors, vol. 12, no. 12, pp. 16 937–16 953, 2012.

[20] S. Hu and G. Zheng, “Driver drowsiness detection with eyelid related parameters by support vector machine,” Expert Systems with Applications, vol. 36, no. 4, pp. 7651–7658, 2009.

[21] M. Ingre, T. A° kerstedt, B. Peters, A. Anund, and G. Kecklund, “Subjective sleepiness, simulated driving performance and blink duration: examining individual differences,” Journal of Sleep Research, vol. 15, no. 1, pp. 47–53, 2006.

[22] F. Abtahi, A. Anund, C. Fors, F. Seoane, and K. Lindecrantz, “Association of drivers’ sleepiness with heart rate variability: A pilot study with drivers on real roads,” EMBEC & NBC 2017, pp. 149–152, 2018.

[23] R. Simons, M. Martens, J. Ramaekers, A. Krul, I. Kl¨opping-Ketelaars, and G. Skopp, “Effects of dexamphetamine with and without alcohol on simulated driving,” Psychopharmacology, vol. 222, pp. 391–399, 2012.

[24] T. Hwang, M. Kim, S. Hong, and K. S. Park, “Driver drowsiness detection using the linear EEG,” 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 4646–4649, 2016.

[25] G. Li, B.-L. Lee, andW.-Y. Chung, “Smartwatch-based wearable EEG system for driver drowsiness detection,” IEEE Sensors Journal, vol. 15, no. 12, pp. 7169–7180, 2015.

[26] 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),” 1st International Conference on Computers, Communications, Signal Processing with Special Track on Biomedical Engineering, pp. 165–168, 2005.

[27] A. A. Hayawi and J.Waleed, “Driver’s drowsiness monitoring and alarming auto-system based on EOG signals,” 2nd International Conference on Engineering Technology and its Applications (IICETA), pp. 214–218, 2019.

[28] G. D. Furman, A. Baharav, C. Cahan, and S. Akselrod, “Early detection of falling asleep at the wheel: A heart rate variability approach,” Computers in Cardiology, pp. 1109–1112, 2008.

[29] R. Bhardwaj, P. Natrajan, and V. Balasubramanian, “Study to determine the effectiveness of deep learning classifiers for ECG based driver fatigue classification,” IEEE 13th International Conference on Industrial and Information Systems (ICIIS), pp. 98–102, 2018.

[30] M. Patel, S. Lal, D. Kavanagh, and P. Rossiter, “Applying neural network analysis on heart rate variability data to assess driver fatigue,” Expert Systems with Applications, vol. 38, no. 6, pp. 7235–7242, 2011.

[31] BITalino (r)evolution Plugged Kit Data Sheet, BITalino, 2016, rev. A.

[32] M. Basner and D. F. Dinges, “Maximizing sensitivity of the psychomotor vigilance test (PVT) to sleep loss,” Sleep, vol. 34, no. 5, pp. 581–591, 05 2011.

[33] S. Loh, N. Lamond, J. Dorrian, G. Roach, and D. Dawson, “The validity of psychomotor vigilance tasks of less than 10-minute duration,” Behavior Research Methods, Instruments, Computers, vol. 36, no. 2, pp. 339–346, 2004.

[34] K. Kaida, M. Takahashi, T. A° kerstedt, A. Nakata, Y. Otsuka, T. Haratani, and K. Fukasawa, “Validation of the Karolinska sleepiness scale against performance and EEG variables,” Clinical Neurophysiology, vol. 117, no. 7, pp. 1574– 1581, 2006.

[35] A. Miley, G. Kecklund, and T. A° kerstedt, “Comparing two versions of the Karolinska sleepiness scale (KSS),” Sleep and Biological Rhythms, vol. 14, no. 3, pp. 257–260, 2016.

[36] F. Shaffer and J. P. Ginsberg, “An overview of heart rate variability metrics and norms,” Frontiers in Public Health, vol. 5, p. 258, 2017.

[37] L. G. Tereshchenko and M. E. Josephson, “Frequency content and characteristics of ventricular conduction,” Journal of electrocardiology, pp. 933–937, 2015.

[38] J. Pan and W. J. Tompkins, “A real-time QRS detection algorithm,” IEEE Transactions on Biomedical Engineering, vol. BME-32, no. 3, pp. 230–236, 1985.

[39] M. Malik, J. T. Bigger, A. J. Camm, R. E. Kleiger, A. Malliani, A. J. Moss, and P. J. Schwartz, “Heart rate variability: Standards of measurement, physiological interpretation, and clinical use,” European Heart Journal, vol. 17, no. 3, pp. 354–381, Mar 1996.

[40] W. G. Bickley, “Piecewise Cubic Interpolation and Two-Point Boundary Problems,” The Computer Journal, vol. 11, no. 2, pp. 206–208, 08 1968.

[41] A. Patanaik, C. K. Kwoh, E. C. Chua, J. J. Gooley, and M. W. Chee, “Classifying vulnerability to sleep deprivation using baseline measures of psychomotor vigilance,” Sleep, vol. 38, no. 5, pp. 723–734, 05 2015.

[42] R. S. M. de Barros, J. I. G. Hidalgo, and D. R. de Lima Cabral, “Wilcoxon rank sum test drift detector,” Neurocomputing, vol. 275, pp. 1954–1963, 2018.

[43] L. Salvati, M. d’Amore, A. Fiorentino, A. Pellegrino, P. Sena, and F. Villecco, “On-road detection of driver fatigue and drowsiness during medium-distance journeys,” Entropy, vol. 23, no. 2, 2021.

[44] “General safety regulation: Overview of the approach for driver drowsiness and attention warning (DDAW),” 2020. [Online].
Available: https://circabc.europa.eu/sd/a/7f9912b5-e180-43ab-b9da-82ddfb1cde8a/ 20200701%20MVWG%20DDAW%20draft%20technical%20req.pdf

[45] M. Peiris, R. Jones, P. Davidson, G. Carroll, P. Parkin, T. Signal, M. van den Berg, and P. Bones, “Identification of vigilance lapses using EEG/EOG by expert human raters,” IEEE Engineering in Medicine and Biology 27th Annual Conference, pp. 5735–5737, 2005.

[46] M. Basner and D. F. Dinges, “An Adaptive-Duration Version of the PVT Accurately Tracks Changes in Psychomotor Vigilance Induced by Sleep Restriction,” Sleep, vol. 35, no. 2, pp. 193–202, 02 2012.

[47] H. Lee, J. Lee, and M. Shin, “Using wearable ECG/PPG sensors for driver drowsiness detection based on distinguishable pattern of recurrence plots,” Electronics, vol. 8, no. 2, 2019.