What is it about?
The objective of NASA's Urban Air Mobility program is to safely perform autonomous flights in complex urban environments. A proposed eXplainable Artificial Intelligence (XAI) model, based on Shapley values, provides interpretability for decision-making in handling off-nominal events. Tested on flight data with and without faults, the model identifies key input features during adverse events, enabling the mission manager to make informed decisions in selecting appropriate Courses of Action (CoA).
Featured Image
Photo by Jametlene Reskp on Unsplash
Why is it important?
Autonomous systems that can explain their decision making processes to a human user provide transparency, accountability, regulatory assistance and human-AI collaboration.
Perspectives
Developing a fully functional, explainable cognitive mission manager is essential to realize the potential of utilizing autonomous flights in complex urban settings under various weather conditions while ensuring safety.
Shivakumar Ranganathan
Read the Original
This page is a summary of: Off-Nominal Event Analysis in Autonomous Flights Based on Explainable Artificial Intelligence, January 2024, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2024-0720.
You can read the full text:
Contributors
The following have contributed to this page
