Project AI/Astro

Science-PI:

    Atri Dutta, Aerospace Engineering, Wichita State University (WSU) 

Co-Investigators:

   James Steck, Aerospace Engineering, WSU 

   Craig McLaughlin, Aerospace Engineering, University of Kansas (KU)  

   Arslan Munir, Computer Science, Kansas State University (KSU)

NASA Technical Monitor: 

   John Dankanich, Chief Technologist, NASA Marshall Space Flight Center 

Industry Collaborator: 

   Pradipto Ghosh, Senior Mission Design and Navigation Engineer, JHU Applied Physics Laboratory  

STEM Partner: 

   William Polite, Director of Equity, Diversity and Accountability, Wichita Public School 

Students (directly funded by project): 

   Amrutha Dasyam, PhD Student, Aerospace Engineering, WSU (2022-Current)

   Adrian Arustei, PhD Student, Aerospace Engineering, WSU (2022-2023) 

   Kyle Messick, PhD Student, Aerospace Engineering, WSU (2022)  

   Yrithu, PhD Student, Aerospace Engineering, WSU (2020-2023) 

   Matthew Chace, MS Student, Aerospace Engineering, WSU (2022-2024) 

   Ella Kreger, UG Student, Aerospace Engineering (2024)  

   Melvin Rafi, PhD Student, Aerospace Engineering (2020) 

   Syed Talha Zaidi, PhD Student, Computer Science, KSU (2022-2024)  

   Ali H. Mughal, MS Student, Computer Science, KSU (2021-2022)   

   Mahmood Azhar Qureshi, Computer Science, KSU (2021)  

   Hayat Ullah, PhD Student, Computer Science, KSU (2022)  

   Charles A. Fry, MS Student, Aerospace Engineering, KU (2021-2023)     

 Other students engaged for research experience: 

   Pardhasai Chadalavada 

   Tanzimul Farabi 

   Ramses Young   

Motivation: Spacecraft trajectory optimization is a critical aspect of space mission analysis. In recent years, there has been an increased interest within NASA in applying machine-learning algorithms to improve the performance of trajectory optimization solvers. Optimization of trajectories for spacecraft employing solarelectric propulsion is a challenging problem because it requires the solution of a nonlinear, non-convex mathematical programming problem. This problem is even more complicated when the spacecraft is located close to a planetary body. First, the low-thrust propulsion system provides a small acceleration relative to the local gravitational acceleration, making the transfer long and complex. Second, the presence of the planets shadow prohibits thrust generation by electric thrusters, thereby making the transfer multi-phase. Third, gravitationally trapped radiation degrades the spacecraft solar array that powers the electric thrusters.

Proposal: The proposed research considers the development of a new, machine-learning assisted optimization tool for on-ground mission design. The automated, fast and robust nature of the proposed methodology makes the tool suitable for onboard implementation as well. The architecture of the proposed software allows for sequential progression of fidelity by incorporating increasingly rigorous force models at different levels of trajectory optimization; this facilitates the improvement of lower-fidelity solutions, while simultaneously managing the computational complexity of the underlying problem in an automated manner. Proposed modular architecture allows for application of the proposed software in two different settings, such as preliminary mission analysis by ground personnel and onboard mission planning. The overall trajectory design is modelled as a two-level process, with the low-level trajectory optimization phase, and a high-level planning that allows for the application of machine learning techniques to trajectory optimization. The proposal considers the following innovations: using dynamical coordinates in trajectory optimization, a modified state observer to estimate unmodeled acceleration, and the use of an artificial neural network for adaptive tuning of planning variables. Additionally, in the context of onboard implementation, we consider data driven updates of the neural networks based on information obtained for sensors. The proposed project will also consider the addition of atmospheric drag models for analysis of aero-capture and atmospheric entry. 

Arustei, A., Dutta, A (2024). Direct Optimization of Low-Thrust Orbit-Raising Maneuvers using Adjoint Sensitivities,. Acta Astronautica. Vol. 219, pp. 965-981.  https://doi.org/10.1016/j.actaastro.2024.03.059  

***The above paper is available for free download through the following link (till June 5, 2024): https://authors.elsevier.com/c/1ixNo_29e80m0u

Pillay, Y., Chace, M., Steck, J., Watkins, J., and Dutta, A (2024). Neuro-adaptive Model Reference Tracking Controller for Cislunar Missions. AIAA Guidance Navigation and Control Conference,
AIAA SciTech Forum, Orlando FL. AIAA 2024-0509. https://doi.org/10.2514/6.2024-0509 

Dutta, A., Arustei, A., Chace, M., Chadalavada, P., Steck,  J., Zaidi, T. & Munir, A (2024). Machine
Learning Assisted Low-Thrust Orbit-Raising: A Comparative Assessment of a Sequential Al-
gorithm and a Deep Reinforcement Learning Approach (2024).  AAS/AIAA Space Flight Mechanics Meeting, AIAA SciTech Forum. Orlando FL. AIAA 2024-1669. https://doi.org/10.2514/6.2024-1669 

Zaidi, A., Chadalavada, P., Ullah, H., Munir, A., and Dutta, A (2023). Cascaded Deep Reinforcement
Learning-Based Multi-Revolution Low-Thrust Spacecraft Orbit-Transfer. IEEE Access,
vol. 11, pp. 82894-82911, https://ieeexplore.ieee.org/document/10207710.  

Fry, C. A (2023). An Exploration of Solar Radio Flux Forecasting Using Long Short-Term Memory Artificial Neural Networks. MS Thesis, University of Kansas.

Dasyam, A., Dutta, A (2023). Artificial Neural Network based Atmospheric Density Model for Aerobraking Trajectory Design. AAS/AIAA Space Flight Mechanics Meeting. Austin TX.

Mughal, A., Chadalavada, P., Munir, A., Dutta, A., & Qureshi, M (2022). Design of deep neural networks for transfer time prediction of spacecraft electric orbit-raising. Elsevier Intelligent Systems with Application. Vol. 15, Article No 200092. https://doi.org/10.1016/j.iswa.2022.200092

Pillay, Y., Chace, M., Steck, J., & Dutta, A (2022). Neural network for predicting unmodelled dynamics in multi-revolution transfers in cis-lunar missions. AAS/AIAA Astrodynamics Specialist Meeting. Charlotte, NC.

Arustei, A., Dutta (2022), A. An adjoint sensitivity method for the sequential low-thrust orbit-raising problem. AAS/AIAA Astrodynamics Specialist Conference. Charlotte NC.

Fry, C., McLaughlin, C (2022). Optimizing Long Short-Term Memory Neural Network to Forecast Solar Radio Flux. AAS/AIAA Astrodynamics Specialist Conference. Charlotte NC.  

Dasyam, A., Chadalavada, P., Fry, C., Dutta, A., & McLaughlin, C (2021). Neural Network Based Estimation of Atmospheric Density during Aerobraking. AAS/AIAA Astrodynamics Specialist Meeting. Held virtually.

Pillay, Y., Chace, M., Messick, K., Steck, J., & Dutta, A (2021). Modified State Observer for Characterization of Unmodeled Dynamics in Cis-lunar Missions. AAS/AIAA Astrodynamics Specialist Meeting. Held virtually.

Chadalavada, P., Dutta, A., & Ghosh, P (2021). An Efficient Algorithm for the Longitude-Targeted Ascent of All-Electric Satellites. AAS/AIAA Space Flight Mechanics Meeting (AIAA Scitech Forum). San Diego CA.

Farabi, T., & Dutta, A. (2021). Artificial Neural Network Based Prediction of Solar Array Degradation during Electric Orbit-Raising. AAS/AIAA Space Flight Mechanics Meeting. Virtual Conference. AAS 21-424.

Lessons Learned with NASA Innovation and Technology Development. Public tak by NASA Chief Technologist John Dankanich, at Wichita State University. January 13, 2022.  

Astronautics Summer Camp, 2022. 

This project is sponsored by NASA EPSCOR CAN program.