Workload analysis of pilot steep turn maneuvers using SR20 aircraft and EEG data

Area of Science:

• Neuroscience
• Aviation Psychology
• Machine Learning in Healthcare

Background:

• Steep turns are critical flight maneuvers requiring significant cognitive resources.
• Understanding the neurophysiological demands of different turn directions is crucial for optimizing pilot training.
• Electroencephalogram (EEG) offers a non-invasive method to capture brain activity related to cognitive workload.

Purpose of the Study:

• To compare the neurophysiological signatures of left versus right steep turns using EEG.
• To leverage machine learning to identify workload-related brain activity patterns.
• To explore potential EEG markers for real-time workload monitoring in aviation.

Main Methods:

• Thirty-seven flight cadets performed left and right steep turns in a flight simulator.
• 32-channel EEG data were recorded and analyzed using sliding windows.
• Extensive features (time, frequency, non-linear domains) were extracted and processed by six machine learning classifiers, with LightGBM showing superior performance.

Main Results:

• Subjective workload (NASA-TLX) was significantly higher during right turns compared to left turns.
• Machine learning models, particularly LightGBM, effectively differentiated between turn types using EEG features.
• Distinct neurophysiological patterns were observed: left turns showed higher high-frequency activity, while right turns exhibited stronger theta/alpha patterns.

Conclusions:

• Directional differences in neurophysiological workload signatures during steep turns are supported by EEG and machine learning analysis.
• Candidate EEG markers were identified for lightweight, real-time workload monitoring.
• Findings can facilitate optimized flight training protocols and enhance overall aviation safety.