An SVM-based identification of acute stress in pilots using objective indicators

Area of Science:

• Aviation Psychology
• Human Factors Engineering
• Aerospace Medicine

Background:

• Sudden flight incidents induce acute stress, potentially compromising aviation safety.
• Objective monitoring of pilot stress is crucial for enhancing flight safety protocols.

Purpose of the Study:

• To investigate physiological and biochemical responses to acute stress in flight cadets during a simulated engine failure.
• To develop and validate a machine learning model for objective pilot stress detection.

Main Methods:

• Simulated engine failure during landing in a C172-G1000 simulator for 37 flight cadets.
• Collected data included state anxiety, heart rate, heart rate variability, salivary cortisol, altitude, and heading.
• A support vector machine model was developed using feature selection, normalization, and hyperparameter tuning.

Main Results:

• Significant differences observed in physiological, biochemical, and behavioral data between stress and non-stress conditions.
• The support vector machine model achieved 86.49% accuracy in distinguishing between stress and non-stress states.
• Identified key indicators for objective stress monitoring in pilots.

Conclusions:

• The study successfully demonstrated a method for objectively monitoring pilot stress levels during critical flight phases.
• Findings support the integration of stress monitoring tools in aviation training and operations.
• Objective stress assessment can contribute to improved pilot performance and flight safety.