Quantitative Evaluation and Factor Analysis of Mental Fatigue During Tractor Operation Based on EEG Sensing

Fatigue during agricultural machinery operation is a critical factor affecting safety and performance, especially under prolonged and harsh working conditions. This study aimed to quantitatively evaluate mental fatigue during tractor operation and to identify its neurophysiological basis. A simulated field operation was conducted, replicating tractor-based crop monitoring under four environmental conditions varying in thermal comfort, noise, and vibration levels. Three healthy male participants performed 2-hour driving tasks under each condition. Subjective (arousal, concentration), behavioral (contact with poles), and physiological (Electroencephalography) indicators were measured. A behavioral index, defined as the onset of a sharp increase in contact errors (fatigue onset), showed temporal correspondence with subjective fatigue reports and was modulated by environmental load. EEG-based regression models using power spectral density features predicted subjective fatigue with high accuracy (R² = 0.81). Feature importance analysis revealed key involvement of brain regions associated with auditory processing, somatosensory integration, attention, and motor control, as well as visual processing and cognitive functions. These findings suggest that fatigue onset reflects multi-domain neural alterations and can be objectively estimated via EEG sensing. The proposed framework contributes to a physiological understanding of mental fatigue and holds potential for application in the design of agricultural machinery that supports operators.