Intelligent Human Operator Mental Fatigue Assessment Method Based on Gaze Movement Monitoring

Sensors, Год журнала: 2024, Номер 24(21), С. 6805 - 6805

Опубликована: Окт. 23, 2024

Modern mental fatigue detection methods include many parameters for evaluation. For example, researchers use human subjective evaluation or driving to assess this condition. Development of a method detecting the functional state is an extremely important task. Despite fact that operator support systems are becoming more and widespread, at moment there no open-source solution can monitor based on eye movement monitoring in real time with high accuracy. Such allows prevention large number potential hazardous situations accidents critical industries (nuclear stations, transport systems, air traffic control). This paper describes developed movements. We our research earlier dataset included captured eye-tracking data operators implemented different tasks during day. In scope method, we propose technique determination most relevant gaze characteristics detection. The includes following machine learning techniques classification: random forest, decision tree, multilayered perceptron. experimental results showed as follows: average velocity within fixation area; curvature trajectory; minimum saccade length; percentage fixations shorter than 150 ms; proportion spent milliseconds. processing using proposed performed time, maximum accuracy (0.85) F1-score (0.80) reached forest method.

Язык: Английский

---

FatigueSense: Multi-Device and Multi-Modal Wearable Sensing for Detecting Mental Fatigue

ACM Transactions on Computing for Healthcare, Год журнала: 2024, Номер unknown

Опубликована: Дек. 21, 2024

Mental fatigue is a crucial aspect that has gained attention across various disciplines due to its impact on overall well-being. While previous research explored the use of wearable devices for detecting mental fatigue, limited investigation been conducted into effectiveness these in different body positions or multi-device setups. To address this, our study utilizes unique public dataset containing over 13 hours sensor data collected 36 sessions, with four (Earable, Chestband, Wristband, and Headband). We propose several machine learning-based approaches assess both psychological physiological levels multimodal environment. Specifically, we introduce device type-specific (trained tested single device) multiple devices) inference tasks. Our findings show models perform well, AUC scores ranging from 0.63 0.69 0.74 0.80 fatigue. The approach shows improved performance (AUC 0.74) 0.81 0.88). Hence, this presents in-depth analysis wearables, demonstrating potential setups are prevalent today’s emerging lifestyles.

Язык: Английский

---

Measured increases in steering entropy may predict when performance will degrade: A driving simulator study

Transportation Research Part F Traffic Psychology and Behaviour, Год журнала: 2022, Номер 91, С. 87 - 94

Опубликована: Окт. 14, 2022

Язык: Английский

---

State space model detection of driving fatigue considering individual differences and time cumulative effect

International Journal of Transportation Science and Technology, Год журнала: 2024, Номер 13, С. 200 - 212

Опубликована: Янв. 21, 2024

Fatigue is an important cause of traffic crashes, and effective fatigue detection models can reduce these crashes. Research has found large differences in fatigued driving performance from driver to driver, as well a significant cumulative effect on given over time. Both sources variation decrease the accuracy systems, but previous studies have not done enough evaluate differences. The purpose this study therefore develop model that considers individual time fatigue. Data lateral position car its lane, steering wheel movement, speed, eye movement were collected 22 drivers using simulator with eye-tracking system. Drivers' subjective scores Karolinska Sleepiness Scale. State space (SSMs) built detect each considering his or her features. As series model, SSM also address fatigue, it does require dataset achieve high levels accuracy. results confirm diversity exist among drivers' performance, so ability take into account driver-specific information suggests more suitable for than use aggregated data. Results show (77.73%) higher artificial neural network (61.37%). advantages accuracy, interpretability, flexible make comprehensive valuable individualized commercial use.

Язык: Английский

---

Transforming fatigue assessment: Smartphone-based system with digitized motor skill tests

International Journal of Medical Informatics, Год журнала: 2023, Номер 177, С. 105152 - 105152

Опубликована: Июль 20, 2023

Язык: Английский

---