疲劳驾驶存在巨大的潜在风险，可能导致严重交通安全事故，而驾驶员状态监测系统是应用程度高且较为有效的驾驶员疲劳监测方法。然而，目前驾驶员状态监测系统较高的误判率导致人机交互体验感下降，及驾驶员对监测系统不信任程度上升，因此驾驶员疲劳监测系统还存在较大提升空间。由于外界环境和工作负荷的差异，驾驶员疲劳分为主动和被动两种状态。基于疲劳的产生和影响机制，本研究认为驾驶员疲劳监测算法应基于不同外界环境和工作负荷而有所调整，因此执行两个实证性研究，探讨主动疲劳和被动疲劳对驾驶员生理和行为指标的影响有何差异。基于以往研究者的探索，本文第一、二个研究问题为探讨不同种类疲劳分别对驾驶员心率和手部运动影响的差异。另外，近年来智能穿戴设备普及度和心率转化技术成熟度不断上升，本文第三、四个研究问题分别为探讨智能手表光电容积信号（PPG光信号）心率采集和陀螺仪手部运动记录功能应用于驾驶员疲劳监测的可行性。本文研究1为在复杂城市道路环境下的主动疲劳驾驶研究，通过智能手表和心率带采集驾驶员心率和手部运动数据。本文研究2在单调高速道路环境下开展，通过智能手表、眼动仪、CAN bus系统采集的数据，探讨被动疲劳对驾驶员心率、眼动、手部运动和驾驶绩效指标的影响。在眼动指标中，本研究添加常用于分心驾驶监测的视觉注视离散程度指标，探讨其对于驾驶员疲劳预测的效果。研究结果发现，主动疲劳研究中，驾驶员心率下降，手部运动无显著变化；被动疲劳研究中，驾驶员心率无显著变化，而手部运动角速度显著下降，且视觉注视离散程度增大。本研究通过模拟导致不同种类疲劳的真实驾驶环境，探讨主动疲劳和被动疲劳对于驾驶员心率、手部运动的影响，并将智能手表运用于数据采集，探讨智能手表作为驾驶员疲劳预测的可行性。在被动疲劳研究中，还增加对眼动指标和驾驶绩效的探索。研究结果不仅能够进一步加深对疲劳复杂机制的理解，还有利于驾驶员状态监测系统技术提升，维护交通安全及友好人机交互体验。
Driver fatigue is a huge potential risk that can lead to serious traffic accidents, and driver monitoring systems (DMS) are a highly effective way to reduce fatigue-related accidents. However, the high false-positive rate of the current DMS has led to the decrease of human-vehicle interaction experience and the increase of the driver's distrust of the monitoring system, so there is still much room for improvement of DMS.The variety of external environment and workload will lead to different kinds of fatigue states, i.e., active fatigue and passive fatigue. Based on the mechanisms of fatigue generation and effects, the current study suggested that DMS algorithms should be adjusted based on different external environments and workloads. Therefore, two empirical studies were performed to investigate the differences in the effects of active and passive fatigue on drivers' physiological and behavioral indicators. Based on previous explorations, the first and second research questions of the current thesis were to explore the differences in the effects of different types of fatigue on drivers' heart rate and hand movement, respectively. In addition, the popularity of smart wearable devices and the maturity of heart rate technology have been rising in recent years; the third and fourth research questions were to explore the feasibility of applying the smart watch Photoplethysmography (PPG)-based heart rate acquisition and the gyroscope hand movement recording function for driver fatigue monitoring. Study 1 in the current thesis was an active fatigue driving study in a complex urban road environment, in which drivers’ heart rate and hand movement data were collected through a smartwatch and a heart rate detector. Study 2 was conducted under a monotonous highway environment to investigate the effects of passive fatigue on drivers’ heart rate, eye movement, hand movement, and driving performance using data collected by a smartwatch, eye tracker, and CAN bus system. Among the eye tracking indicators, the current study added the visual gaze dispersion, which has been commonly used for driver distraction monitoring, to investigate its effectiveness for driver fatigue monitoring. The results showed that in the active fatigue study, drivers’ heart rate decreased and there was no significant change in drivers’ hand movement; in the passive fatigue study, no significant change was found in heart rate, while the angular velocity of hand movement decreased and the visual gaze dispersion increased.The current thesis investigated the effects of active and passive fatigue on drivers’ heart rate and hand movement in real driving environments that lead to different types of fatigue, and explored the feasibility of smart watches as driver fatigue monitoring by applying smart watches for data collection. In the passive fatigue experiment, eye tracking indicators and driving dynamics were also discussed. The results of the current thesis not only further deepen the understanding of the complex mechanisms of fatigue, but also contribute to the technical improvement of DMS, maintain traffic safety and improve friendly human-vehicle interaction experience.
