Periodic Reporting for period 2 - THOR (Towards Higher Levels of Autonomy and Robustness in Space Operations through Uncertainty Management and Quantification)
Periodo di rendicontazione: 2023-05-01 al 2024-04-30
From a societal perspective, THOR is mainly involved in advancing deep space exploration. In particular, it focuses on asteroids, which are of great scientific interest since they are remnants of the early Solar System period. Moreover, certain types of asteroids contain valuable resources such as water, metals, and minerals. These resources could be mined and utilised in space exploration endeavours, potentially enabling long-duration space missions through resource utilisation. The THOR project contributes to these areas by developing methods to characterize gravity around asteroids while in flight. Asteroid details are hardly observable by Earth-based sensors; thus, they have to be characterised in situ.
THOR project objectives can be succinctly resumed in:
I. How can we solve the asteroid gravimetry problem using on-board optical navigation measurements?.
II. How can we fuse physics-founded gravity models with neural network models?.
Subsequently, during the return phase at U. Sevilla, THOR aimed to refine the previously optimized mascon gravity models by integrating them with neural networks. This approach amalgamates the strengths of a physics-founded model, which extrapolates effectively in data-sparse regions, with the high accuracy provided by neural networks within the dataset area. The outcome is a highly accurate yet lightweight fused gravity model tailored for asteroids.
THOR research outputs are:
J.C. Sanchez, H. Schaub, "Simultaneous navigation and mascon gravity estimation around small bodies," Acta Astronautica, vol. 213, pp. 725-740, 2023.
J.C. Sanchez, J.R. Martin, "Investigating the fusion of mascon and neural network gravity models", submitted to AAS/AIAA Astrodynamics Specialist Conference, 2024.
J.C. Sanchez, H. Schaub, " Small body navigation and gravity estimation using Kalman filter and least-squares fitting", 33rd AAS/AIAA Space Flight Mechanics Meeting, Austin, USA, 2023.
J.C. Sanchez, H. Schaub, "Semi-autonomous navigation and gravity estimation around small bodies", 2nd International Stardust Conference, Noordwijk, The Netherlands, 2022.
THOR dissemination comprises:
https://cafeconciencia.fundaciondescubre.es/noticias/estudiantes-de-bachillerato-se-toman-un-cafe-con-ciencia-aeroespacial-para-celebrar-la-semana-del-espacio/
https://lanochedelosinvestigadores.fundaciondescubre.es/investigador/julio-cesar-sanchez-merino/
https://www.us.es/actualidad-de-la-us/11-investigaciones-de-la-us-reciben-ayudas-marie-curie
Regarding the topic of gravity field modeling, THOR has explored the concept of integrating robust physical models with data-driven neural networks. Researchers have devoted extensive efforts, with ongoing progress, to develop accurate yet lightweight gravity models for on-board autonomy. In THOR, pre-trained mascon models have been combined with physics-informed neural networks (PINNs). Mascon models offer a precise initial approximation to the asteroid's gravity field and effectively extrapolate the high-altitude profile where data is lacking. However, their learning capacity is constrained by the finite number of masses used, and errors may arise from mass placement near the surface. Recent advancements in PINN gravity models, demonstrated by J. Martin and H. Schaub, have showcased their capability to approximate highly complex patterns, such as surface gravity. Nonetheless, these authors observed neural networks' limitations in extrapolating patterns beyond data bounds, prompting the inclusion of low-fidelity analytical models (e.g. point mass) to prevent divergence. THOR proposes to enhance the low-fidelity analytic component with pre-trained mascon models, which exhibit accuracy except in areas very close to the surface. This approach enables continued learning of complex patterns by the PINN through the introduction of new basis functions. Currently, these models have been implemented in the open-source Basilisk astrodynamics framework to benefit the broader astrodynamics community. Preliminary results indicate a one-order-of-magnitude decrease in 1-day propagation errors for orbits around asteroid Eros with the fused model. We anticipate completing and submitting a journal version of this work after summer 2024.