ABSTRAK

Improving aircraft design relies on accurate aerodynamic modeling for predictingflight characteristics, wherein system identification and parameter estimationplay a pivotal role. While the existing method demonstrates adequacy inachieving the desired outcomes, this research explores the potential of applyinga deep learning approach for parameter estimation. This research assesses theefficacy of Deep Learning in aircraft aerodynamic model identification, comparingit to the established Least Squares method. Utilizing flight test datafrom the ATTAS research aircraft, the study demonstrates that Deep Learningeffectively captures trends and reduces fluctuations. While all models showgood predictive capability, slight discrepancies with Least Square estimatesand limitations related to data variability and outliers were observed. Futureresearch should focus on overcoming these challenges to enhance the precisionof Deep Learning models for parameter estimation.