An in-flight multimodal data collection method for assessing pilot cognitive states and performance in general...

Area of Science:

• Aviation Psychology
• Human Factors Engineering
• Biomedical Engineering

Background:

• Pilot cognitive states (workload, stress, situation awareness) are critical for aviation safety.
• Current flight simulators often lack ecological validity for replicating real-world pilot cognitive states.
• General aviation, particularly the Cessna 172, is a key area for pilot training and safety research.

Purpose of the Study:

• To present a novel in-flight data collection methodology for general aviation.
• To create a multimodal dataset combining physiological, flight, and self-reported data.
• To support human factors research and applications in pilot training and aviation safety.

Main Methods:

• Utilized a Cessna 172 aircraft for data collection.
• Employed wearable physiological sensors (EEG, ECG, EDA, body temperature) and eye-tracking glasses.
• Integrated ADS-B flight recorder data and instructor-rated performance.
• Developed procedures for sensor setup, flight task design, and data synchronization.

Main Results:

• Collected 20 complete multimodal datasets from 25 participants after data cleaning.
• Demonstrated the feasibility of in-flight multimodal data acquisition.
• Established a foundation for statistical and machine learning analyses of pilot cognitive states.

Conclusions:

• The presented methodology offers high ecological validity for studying pilot cognitive states.
• The dataset has significant practical value for pilot training, performance evaluation, and safety management.
• Further data collection will expand the dataset's utility for advanced research on pilot behavior and performance.