Machine Learning for Phase of Flight Detection

The safety, operational effectiveness, and training aspects of aviation are all impacted by accurate flight phase identification; as such machine learning approaches have been developed for this purpose. This study first provides a scientometric and conceptual overview of literature that utilizes machine learning methods for flight phase detection. The scientometric approach assesses publishing trends and the evolution of research subjects, to find important papers and large research centers. The conceptual study shows how machine learning has been modified to meet the needs of scientists by addressing its theoretical foundation and practical applications. Further, to determine the efficient methods for identification of crucial flying phases of an aircraft, namely straight and level phase and turn phase, multiple unsupervised clustering algorithms are employed on flight data simulated through Monte Carlo Simulations and on real flight data. The performance of clustering models is evaluated, and it is found that Gaussian mixture model (GMM) outperforms all other models. Additionally, research gaps are identified and prospective routes for future research are offered.