Members of LATTICE Receive Best Paper Award at joint FAA-EUROCONTROL ICRAT Conference

Assistant Professor Max Li, along with recent U-M alumni Sinan Abdulhak and other members of the Laboratory for Air Transportation, Infrastructure, and Connected Environments (LATTICE), received a best paper award at the International Conference on Research in Air Transportation (ICRAT). The awarded paper titled, “ChatATC: Large Language Model-Driven Conversational Agents for Supporting Strategic Air Traffic Flow Management,” covered large language models and their potential application to the current aviation infrastructure.

With Generative AI and Large Language Models (LLMs) being a “hot topic” in a variety of new applications worldwide, Abdulhak began highlighting the potential benefits of this software in the aviation industry. Using this research, the team has focused on finding responsible and practical uses for LLMs to help air traffic managers accomplish simple and repetitive tasks more efficiently.

Working closely with Professor Li, Abdulhak and team detailed a strategic application of the ChatATC conversational agent, covering the use of Ground Delay Programs (GDP) which are measures used to slow the rate of arrivals into an airport due to some constraints such as poor weather. Using this model, Traffic Managers will have the ability to rapidly summarize thousands of historical air traffic events and formulate decisions for future events more quickly.

“Air Traffic Managers often issue GDPs for similar reasons, but currently don’t have the resources to quickly see how a repetitive issue was addressed in the past,” Abdulhak explained. “We wanted to see if LLMs could address this by rapidly generating tailored summaries of historical GDPs.”

Throughout this project, the team collaborated with a National Traffic Management Officer from the FAA to gain real world experiences and information. They plan to use this information and carefully think about how this new technology can practically address some of the user burdens, allowing them to execute their job more efficiently.

Going forward, the team plans to continue exploring ways to make the ChatATC’s summaries of historical GDPs more robust through further analysis and collaboration with natural language processing researchers and FAA stakeholders. To do this, they will continue testing the conversational agent’s responses with FAA Traffic Managers.

“We are also planning on collaborating with software developers to build out a graphical user interface that will facilitate the interaction between traffic managers and ChatATC,” Abdulhak stated.

About the conference:

ICRAT is globally known as being a highly competitive air transportation research conference, jointly organized by the FAA and EUROCONTROL, which are responsible for the safe and efficient operation of air traffic across US and European airspace. The ICRAT forum provided a platform for young researchers and mentors in air transportation to share their work, expand their professional network, practice scientific dissemination, gain inspiration from others while discovering new challenges confronting society.

This year’s biennial conference, which alternates with the USA/Europe Air Traffic Management (ATM) Research and Development (R&D) Seminar, was hosted at Nanyang Technological University in Singapore, marking a significant milestone of being the first time it has been held in the Asia-Pacific region. This new venue opportunity showcased the rising demand for global air travel and the desire for intercontinental collaborations.

Abdulhak further explained the significance behind attending the conference by stating, “It was also the perfect venue to identify potential collaborations with researchers from other universities and organizations who are conducting research that is relevant to our work. Following my presentation on ChatATC, we received great feedback from participants and industry stakeholders on potential future research directions, and were approached by multiple researchers and FAA stakeholders for potential collaborations to build on our research.”

About the authors:

Professor Li is no stranger to ICRAT recognition, having won two best paper awards in 2022. This year, Li participated in multiple paper sessions throughout the event, was also on the Program Committee and led session chairs on both Tuesday, July 2 and Wednesday, July 3.

Abdulhak, a recent alumni from the University of Michigan Industrial and Operations Engineering (IOE) program, began research on the ChatATC project as part of his Directed Study with Professor Li for his IOE Senior Capstone.

“I could’ve never imagined the platform that I would be able to disseminate our work on. Receiving the best paper award at this conference shows not only the great interest in our work from key aviation stakeholders, but also validates the approach we used to execute our research. As someone who is so passionate about commercial air travel, emerging technologies, and designing great user experiences, this research represents the culmination of all these interests and highlights the practical application of our work. I’m so thrilled to be tackling this research with talented co-authors from Stanford University (Dr. Karthik Gopalakrishnan, now at Tesla) and the FAA (Wayne Hubbard), and grateful to have the guidance and mentorship from Professor Max Li. Lastly, as someone who always dreamed of making commercial air travel more efficient and enjoyable but worried about the limited opportunities to make an impact, this conference showed me the significant demand for user-centric innovations in the aviation industry, compounded by the rise in air travel over the coming decades.”

Authors of the winning paper include: Sinan Abdulhak, Wayne Hubbard, Karthik Gopalakrishnan and Max Li