A Fuzzy-Enhanced Explainable AI Framework for Flight Continuous Descent Operations Classification

Continuous Descent Operations (CDO) offer significant environmental and operational benefits: The study confirms that CDO—smooth descents without level-offs—reduces fuel burn, emissions, and noise while improving efficiency and passenger comfort, demonstrating the practical value of optimizing this flight phase.

Explainability is essential for aviation AI systems: The research emphasizes that transparency in AI decision-making is "critical for safety and stakeholder trust" in aviation, addressing a gap where many existing trajectory optimization methods lack the interpretability that the industry requires.

A small set of key factors drive CDO performance: Using SHAP analysis, the study identified that just three features—average descent rate, number of descent segments, and average change in descent angle—were the most influential factors in determining whether flights adhered to CDO procedures, suggesting that focusing on these metrics could improve compliance.

Fuzzy logic makes AI recommendations more actionable for pilots and operators: By converting machine learning insights into fuzzy rule-based classifiers, the FEXAI framework produced "meaningful, human-readable rules" that operational users can understand and apply, rather than opaque algorithmic outputs.

High accuracy with interpretability is achievable: All models achieved classification accuracies above 90% while maintaining explainability, demonstrating that aviation systems don't have to choose between predictive performance and transparency.