A Data-Driven Approach using Machine Learning to Enable Real-Time Flight Path Planning

As aviation traffic continues to grow, most airlines are concerned about flight delays, which increase operating costs for the airlines. Since most delays are caused by weather, pilots and flight dispatchers typically gather all available weather information prior to departure to create an efficient and safe flight plan. However, they may have to perform in-flight re- planning because weather information can significantly change after the original flight plan is created. One potential issue is that weather forecasts being currently used in the aviation industry may provide relatively unreliable information and are not accessible fast enough so that it challenges pilots to perform in-flight re-planning more accurately and frequently. In this paper, we propose a data-driven approach that uses an unsupervised machine learning technique to provide a more reliable and up-to-date area of convective weather. To evaluate the proposed methodology, we collect the American Airlines flight (AA1300) information and actual weather-related data on October 6th, 2019. Preliminary results show that the proposed methodology provides a better picture of the nearby convective weather activity compared to the most well-known convective weather product.