A comprehensive flight performance evaluation model based on flight parameters with comparison to subjective and AI...

Area of Science:

• Aviation Psychology
• Flight Simulation
• Performance Metrics

Background:

• Developing robust flight performance evaluation models is crucial for effective pilot training.
• Existing models may not adequately cover diverse flight conditions, including emergencies.
• Objective and consistent performance assessment remains a challenge in aviation training.

Purpose of the Study:

• To create a comprehensive flight performance evaluation model using flight parameters.
• To ensure the model's applicability across normal and abnormal flight scenarios.
• To validate the model's reliability against expert and AI-based assessments.

Main Methods:

• Utilized a Cessna-172 simulator for normal and single-engine failure emergency flights with 37 pilots.
• Segmented flights into distinct phases (takeoff to landing, including emergency).
• Employed the analytic hierarchy process for weighting flight phases and metrics; incorporated instructor and ChatGPT evaluations.

Main Results:

• Demonstrated good consistency between the model, flight instructors, and ChatGPT-generated scores.
• Found significant correlations among all evaluation sources, validating the model's reliability.
• The model effectively assessed performance across different flight phases and conditions.

Conclusions:

• The developed flight performance evaluation model is reliable and accurate.
• The model shows potential for practical application in real-world flight training and operations.
• AI tools like ChatGPT can serve as supplementary evaluation sources in aviation training.