Comparative analysis of Neural Networks approaches for Initial Orbit Determination of Near-Earth Objects
Abstract
Initial Orbit Determination (IOD) from Very Short Arcs (VSAs) remains one of the most challenging open problems in asteroid surveillance and celestial mechanics. Classical methods require angular observations spanning a sufficient fraction of the orbit, and become ill-conditioned or fail outright when only a single observing night is available, as is the case for most newly discovered Near-Earth Objects (NEOs). We present a comparative analysis of two Neural Network (NN) models that attack the ranging problem directly: both ingest a triplet of time-tagged angular measurements (ti,αi,δi), i=1,2,3, from a single VSA and predict the geocentric range and range-rate (ρ,ρ̇), thus completing the orbital state vector. The first is a Multi-Layer Perceptron trained on a purely data-driven objective; the second augments it with a physics-informed loss. Both are trained, under an object-disjoint partition, on a sample of the 568 127 VSAs of 39 031 real NEOs available in the NEODyS-2 catalogue. Each model returns an estimate on every arc by construction, so we assess instead whether that estimate is dynamically admissible: on a held-out test set of 56 773 arcs the physics-informed model places 85.1% of its predictions inside the admissible region, against 76.5% for the data-driven baseline, whereas Gauss's and Laplace's methods return a solution on only 45.0% and 47.3% of the same arcs and collapse onto the degenerate root in about nine of those cases out of ten. A stratified analysis over proper motion and true range shows that the physics-informed objective is not uniformly superior to the baseline: it trades a longer error tail for admissibility and for a marked advantage on the fast, nearby arcs that are operationally the most relevant.
Related
- Metaheuristic physics-informed extreme-learning machine for initial orbit determination of near-Earth asteroids from too-short arcs
- Measuring the Age and Initial Composition of Interstellar Objects
- Near-infrared periodicity in the SPICY catalog. True young stellar objects versus contaminants
- Investigating the formation of small Solar System objects using stellar occultations by satellites: present, future and its use to update satellite orbits
- Beneath the surface: >85% of z>5.9 quasi-stellar objects in massive host galaxies are UV-faint
BibTeX
@misc{geroni2026comparative,
title = {{Comparative analysis of Neural Networks approaches for Initial Orbit Determination of Near-Earth Objects}},
author = {Francesco Geroni and Roberto Paoli and Riccardo Massidda and Giacomo Tommei},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.22020},
eprint = {2609.22020},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.22020},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Comparative analysis of Neural Networks approaches for Initial Orbit Determination of Near-Earth Objects
AU - Francesco Geroni
AU - Roberto Paoli
AU - Riccardo Massidda
AU - Giacomo Tommei
PY - 2026
DA - 2026/09/18
PB - arXiv
DO - 10.48550/arxiv.2609.22020
UR - https://arxiv.org/abs/2609.22020
AB - Initial Orbit Determination (IOD) from Very Short Arcs (VSAs) remains one of the most challenging open problems in asteroid surveillance and celestial mechanics. Classical methods require angular observations spanning a sufficient fraction of the orbit, and become ill-conditioned or fail outright when only a single observing night is available, as is the case for most newly discovered Near-Earth Objects (NEOs). We present a comparative analysis of two Neural Network (NN) models that attack the ranging problem directly: both ingest a triplet of time-tagged angular measurements (t_i,α_i,δ_i), i=1,2,3, from a single VSA and predict the geocentric range and range-rate (ρ,ρ̇), thus completing the orbital state vector. The first is a Multi-Layer Perceptron trained on a purely data-driven objective; the second augments it with a physics-informed loss. Both are trained, under an object-disjoint partition, on a sample of the 568 127 VSAs of 39 031 real NEOs available in the NEODyS-2 catalogue. Each model returns an estimate on every arc by construction, so we assess instead whether that estimate is dynamically admissible: on a held-out test set of 56 773 arcs the physics-informed model places 85.1% of its predictions inside the admissible region, against 76.5% for the data-driven baseline, whereas Gauss's and Laplace's methods return a solution on only 45.0% and 47.3% of the same arcs and collapse onto the degenerate root in about nine of those cases out of ten. A stratified analysis over proper motion and true range shows that the physics-informed objective is not uniformly superior to the baseline: it trades a longer error tail for admissibility and for a marked advantage on the fast, nearby arcs that are operationally the most relevant.
ER -