Sungmoon Choi, Ph.D.Iowa State University, Ames, Iowa
Friday, April 10th, 2025, 2:00 - 3:00 PM
Jett Hall, room 109
Title: Seminar Title: Machine Learning-Aided Tour Design and Angles-Only Navigation for Deep-Space Missions
ABSTRACT: With the successful launch of Artemis II, a crewed lunar mission, the 2023–2032 Planetary Science and Astrobiology Decadal Survey identifies a Uranus Orbiter and Probe as the highest-priority new flagship mission and the Enceladus Orbilander as the second-highest priority. These priorities underscore the importance of deep-space missions for advancing our understanding of ice-giant systems and ocean-world habitability. To enable such missions, efficient tour design and onboard navigation are essential, and this talk addresses key challenges in both areas. The first part focuses on mission design. Multi-leg trajectory design and flyby tour optimization are computationally demanding because the trade space grows combinatorially as the number of flyby increases. In addition, each candidate sequence exhibits a distinct cost function topology, making the evaluation of many possible sequences expensive and requiring numerous optimization runs. To address these challenges, the talk first presents a rapid transfer-arc computation method based on a deep neural network and then introduces a novel genetic algorithm for efficiently generating both optimal and suboptimal flyby sequences. It also presents a reinforcement learning-based approach for sequential mission cost estimation that evaluates one trajectory leg at a time without requiring a fully constructed end-to-end trajectory. The second part introduces a geometric approach to angles-only navigation. Unlike conventional methods that depend on a priori attitude information from a star tracker, the proposed approach enables the simultaneous determination of position and attitude without prior attitude knowledge. Within this framework, the navigation problem is formulated as a root-finding problem. The method is further extended to initial orbit determination and state estimation using techniques from complex analysis, including the argument principle and Cauchy’s integral formula.
BIO: Sungmoon Choi holds a B.S. in Mechanical Engineering and two M.S. degrees in Astronomy and Aerospace Engineering. He recently defended his Ph.D. dissertation at Iowa State University, majoring in Aerospace Engineering with a minor in Computer Science. His doctoral studies were supported through the Korean Government Scholarship Program for Study Overseas. During his Ph.D. study, he interned at NASA’s Jet Propulsion Laboratory for over two and a half years, where he worked on AI/ML-based tour design and trajectory optimization and continues to collaborate on related research. Dr. Choi’s research interests include mission design and navigation, spacecraft guidance, navigation, and control (GNC), and machine learning applications in astrodynamics..
