Project Description
The AVS Lab is investigating how advances in machine learning can be used to enhance spacecraft navigation. Specifically, this work will study how Physics-Informed Neural Networks, Neural ODEs, and Lagrangian Networks can enable greater science returns for upcoming interplanetary missions like Lucy, Europa Clipper, and Psyche. These research studies will include training advanced machine learning systems to learn complex gravity and dynamics models, discover new families of orbits from which spacecraft can conduct science, and apply reinforcement learning for spacecraft control. This experience will serve as a strong introduction into advanced spaceflight engineering and offer many opportunities to gain experience working on cutting-edge, real-world machine learning systems.
We are seeking an undergraduate student interested in machine learning systems and astrodynamics. The student should be comfortable with differential equations, basic orbital mechanics, and Python. Please reach out to John Martin if interested.
Special Requirements
- Undergraduate majoring in Aerospace Engineering, Computer Science, or Physics
- Experience working in Python and with moderate-to-large size software projects
- Basic understanding of machine learning and orbital mechanics
- Able to commit to 5-10 hours per week
- (Recommended/Optional) Experience with Tensorflow/Keras
Contact
- Hanspeter Schaub (faculty)
- John Martin (graduate student)