What is it about?
Guidance, navigation, and control (GN&C) becomes challenging in hypersonic flight due to difficulties in measuring the aircraft condition. Temperatures are so high that materials melting and chemical reactions on the outside surface of the vehicle become an issue. We evaluate here how using the vehicle's deformation can be used to recover its aerodynamic state by applying machine learning, and, specifically how we can best generate training data to enable this.
Featured Image
Photo by Hermeus on Unsplash
Why is it important?
We show a cost-effective way of generating the training data based on state-of-the-art modeling techniques. We also consider how the neural network architecture can change according to the type of data that are available.
Perspectives
The problem addressed here felt like an alluring challenge to research on, as it required the integration of such diverse technical disciplines to tackle its complexity.
Ana Cristine Meinicke
University of Michigan
Read the Original
This page is a summary of: On the Efficient Generation of Training Data for High-Speed Aerodynamic State Recovery Through Strain-Load Neural Network Model, January 2024, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2024-1047.
You can read the full text:
Resources
Contributors
The following have contributed to this page
