Related Experiment Video
Updated: May 14, 2025
07:48
Eye Tracking During A Complex Aviation Task For Insights Into Information Processing
Published on: April 4, 2025
158
Situational Awareness Prediction for Remote Tower Controllers Based on Eye-Tracking and Heart Rate Variability Data
Weijun Pan1, Ruihan Liang2, Yuhao Wang2
1Flight Technology and Flight Safety Research Base of the Civil Aviation Administration of China, Civil Aviation Flight University of China, Guanghan 618307, China.
Sensors (Basel, Switzerland)
|April 12, 2025
Summary
Remote tower technology enhances air traffic control efficiency. This study uses eye-tracking and heart rate data to accurately predict air traffic controller situational awareness (SA) in simulated remote towers.
Area of Science:
- Human Factors
- Aerospace Engineering
- Cognitive Science
Background:
- Remote tower technology offers cost savings for airports but introduces challenges to air traffic controller situational awareness (SA).
- Assessing SA in digital environments is crucial for maintaining aviation safety and operational efficiency.
- Existing methods for SA assessment may not fully capture the complexities of remote tower operations.
Purpose of the Study:
- To develop and validate a method for assessing and predicting air traffic controller SA in remote tower environments.
- To investigate the relationship between physiological data and SA in simulated air traffic control scenarios.
- To provide a theoretical foundation for understanding the impact of physiological states on controller SA.
Main Methods:
- Collected eye-tracking (ET) and heart rate variability (HRV) data from participants in a remote tower simulation.
- Annotated a dataset using the Scenario Presentation Assessment Method (SPAM) with probe questions aligned to SA hierarchy and task flow.
- Trained a LightGBM model optimized by Tree-structured Parzen Estimator (TPE) on 25 ET and HRV features, using SHapley Additive exPlanations (SHAP) for interpretation.
Main Results:
- The TPE-LightGBM model demonstrated strong predictive performance for controller SA.
- Achieved Root Mean Square Error (RMSE) of 0.0909, Mean Absolute Error (MAE) of 0.0730, and adjusted R-squared of 0.7845.
- SHAP analysis provided insights into the contribution of different physiological features to SA prediction.
Conclusions:
- The developed TPE-LightGBM model offers an effective approach for real-time SA assessment in remote tower operations.
- Physiological data (ET and HRV) are significant indicators of controller SA in remote tower environments.
- This research contributes to a better understanding of human factors in advanced air traffic control systems.
