Junghyun (Andy) Kim
(Advisor: Prof. Dimitri Mavris]

will defend a doctoral thesis entitled,

Data-Driven Approach using Machine Learning for Real-Time Flight Path Optimization

On

Monday, March 29 at 12:00 p.m.
https://bluejeans.com/995807248

Abstract

While the COVID-19 pandemic has greatly impacted the number of flights and passengers, flight delay-related challenges may reappear soon as aviation traffic continues to rebound from the pandemic. Since flight delays are primarily caused by weather, airlines typically gather all available weather information before departure to generate flight routes that avoid hazardous weather while minimizing operating expenditures. However, pilots potentially have to perform in-flight re-planning as weather information can significantly change after original flight plans are created.

One potential issue is that current in-flight re-planning systems are not fully automated; thus, pilots today perform some portions of the in-flight activities manually. The manual decision-making process may not be currently an issue; however, the advent of new communication systems will bring more information into the cockpit. This has the potential to significantly increase the workload of pilots especially if they must consider a large volume of information, leading to potential safety issues in the future. Another potential issue is that weather forecasts used for current in-flight re-planning systems provide relatively unreliable information and are not always accessible in real-time.

This research attempts to resolve the aforementioned potential issues by developing a framework that automatically performs in-flight re-planning continuously in real-time with the latest weather information sets available. This study specifically develops 1) a supervised machine learning-based wind prediction model to obtain continuous wind information, 2) an unsupervised machine learning-based short-term (i.e., every 10 minutes) convective weather prediction model to define reliable and up-to-date areas of convective weather, and 3) a designated points-based flight path optimization framework that combines the A* search algorithm with an instance-based learning algorithm to find an optimal flight path.

The intent of this research is to provide an automated framework with two use-cases in mind: 1) to help flight dispatchers at major airlines in the U.S. by providing the capability to continuously re-route flights in real-time using the latest weather information and 2) to alleviate the cockpit workload of pilots employed by small airline operators (e.g., private business jets) in the U.S. that do not necessarily have flight dispatchers but rather ask the pilots to generate and update new flight plans.

As a part of this research, a set of case studies is constructed as a proof-of-concept. In addition to the case studies, statistical analyses are performed using 43 real flights (e.g., American Airlines) to prove potential benefits and applicability. The results indicate that the framework developed by this research generates flight routes that reduce flight time by up to two percent in most cases. The outcome of this research establishes not only an automated framework that enables the airlines to perform real-flight path optimization in the contiguous U.S. more accurately and frequently but also provides a basis for optimizing flight routes for all categories of airplanes such as private business jets.

Committee

·         Prof. Dimitri Mavris – School of Aerospace Engineering (advisor)

·         Prof. Daniel Schrage – School of Aerospace Engineering

·         Prof. Polo Chau – School of Computational Science and Engineering

·         Prof. Chao Zhang – School of Computational Science and Engineering

·         Dr. Simon Briceno – Jaunt Air Mobility