PHYSICS-INFORMED CONDITIONAL NORMALIZING FLOWS FOR ANGLES-ONLY CISLUNAR ORBIT DETERMINATION
Walther Litteri, Massimiliano Vasile · 2026
Read PHYSICS-INFORMED CONDITIONAL NORMALIZING FLOWS FOR ANGLES-ONLY CISLUNAR ORBIT DETERMINATION by Walther Litteri, Massimiliano Vasile (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-30936.txt.
Introduction
Walther Litteri and Massimiliano Vasile bring generative modelling to orbit determination around the Moon. Their task is angles-only tracking in cislunar space: from a short arc of direction measurements alone, infer the likely starting state of a spacecraft. A conditional normalising flow learns the density. It is trained on perturbed observations for Near Rectilinear Halo Orbits, the elongated lunar orbits used by current missions. The physics of the motion informs what it learns. The density can hold several distinct candidate states at once.
Given fresh measurements, the flow is sampled to produce state hypotheses that fit both the data and the physics. A least-squares step then polishes them. The authors present the estimates as a strong warm start for classical algorithms.
Catalog record
| CREATOR | Walther Litteri, Massimiliano Vasile |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.30936
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
PHYSICS-INFORMED CONDITIONAL NORMALIZING FLOWS FOR ANGLES-ONLY CISLUNAR ORBIT DETERMINATION Walther Litteri, Massimiliano Vasile Abstract Generative Astrodynamics is advanced in this work by extending generative modelling to an orbit determination problem in the cislunar environment. The task is formulated as conditional density estimation, aiming to infer the probability distribution of the initial state from angles-only measurements over short observation arcs. A normalising flow is trained on perturbed topocentric observations from Near Rectilinear Halo Orbits, enabling a flexible and potentially multimodal posterior representation. Given new measurements, the learned density is sampled to generate statistically consistent and physics-informed state hypotheses. These estimates are refined via nonlinear least-squares minimisation, providing a competitive warm start for classical algorithms. \PaperNumber 26-875 1 Introduction The problem of recovering the orbit followed by a body in space is arguably one of the first scientific problems ever formulated. With the advent of space exploration and the growing exploitation of the space environment, this problem has also become central to the engineering community, which requires reliable methods to estimate the state of a spacecraft from observational data. This field of research is known as Orbit Determination (OD). The literature further distinguishes between different operational scenarios depending on the phase of the mission [ 14 ] . In particular, when orbit estimation is performed at the very beginning of operations, e.g., immediately after launch or following a critical recovery phase, the problem is referred to as Initial Orbit Determination (IOD). A key difference between standard OD and IOD lies in the availability of prior information. In routine OD operations, additional data may be available from previously estimated states or measurements provided by onboard instruments. When such information is unavailable, as in the case of newly launched spacecraft or uncooperative objects, the IOD problem becomes significantly more challenging. Over the past decades, interest in space operations in the cislunar environment has increased considerably. The presence of water ice in permanently shadowed craters near the lunar south pole makes the Moon an attractive location for future scientific and commercial activities [ 1 ] . The Artemis programme represents the most recent international effort aimed at returning humans to the lunar vicinity, with the crewed lunar flyby mission Artemis II now scheduled for April 2026 [ 3 ] . In the original Artemis architecture, the Gateway lunar orbital outpost was conceived as a strategic counterpart to the Earth‑orbiting International Space Station, designed to support subsequent surface missions, including Artemis IV and V. However, following a comprehensive restructuring of the Artemis programme announced in early 2026, the role and priority of the Gateway have become significantly less certain, with current plans shifting towards a more direct and sustained human presence on the lunar surface. Operating in the cislunar environment introduces several additional challenges compared to Earth-orbiting missions. The large Earth–Moon distance complicates communication and synchronisation, introducing significant time delays between command transmission and reception. Moreover, the intrinsic dynamical instability and chaotic behaviour that characterise many cislunar trajectories further increase operational complexity, reducing the margin for allowable errors and demanding higher levels of precision and accuracy. Orbit Determination in cislunar space is further complicated by the absence of a dedicated Global Navigation Satellite System (GNSS) capable of supporting navigation in the region, as well as by limited communication opportunities that restrict the availability of frequent tracking measurements. Traditional approaches to orbit determination typically rely on correction schemes, such as minimising the nonlinear mean squared error in a statistical framework. The Extended Kalman Filter (EKF) integrates knowledge of the dynamics to improve estimation accuracy. More recently, machine learning and statistical learning techniques have been explored to model unmodelled forces through nonlinear latent force models, which act as filters for uncorrelated Gaussian noise [ 6 , 2 ] . Physics-Informed Neural Networks (PINNs) are neural networks trained with an additional term enforcing the underlying physical laws [ 16 , 17 ] . In the context of OD, PINNs have been used to approximate orbit propagation and measurement models in the cislunar environment, thereby facilitating the solution of the inverse problem [ 13 ] . Generative Artificial Intelligence (AI) refers to deep learning strategies used to model probability distributions and to sample from them. Applications of this framework span a variety of domains, from the generation of text, images, and audio [ 10 ] . In astrodynamics, within the emerging field of Generative Astrodynamics, generative methods based on Variational Autoencoders have been successfully applied to analyse and generate periodic orbits in the Circular-Restricted Three-Body Problem (CR3BP) [ 8 , 9 ] , while Normalizing Flow models have been employed for the identification and characterisation of equilibrium solutions in the N-body problem [ 18 ] . This work presents a novel application of Generative Astrodynamics to the solution of an orbit determination problem in the cislunar environment. Topocentric states corresponding to an observation arc are modelled as a distribution conditioned on angles-only observations. A Normalizing Flow model is trained on this distribution, and sampling conditioned on observations provides the estimated states, which are further refined using traditional methods. It will be shown how the inclusion of an additional term in the loss, based on the propagation of the trajectory, helps regularise the physics loss, yielding more accurate results. The paper is structured as follows. First, an overview of the problem is provided, defining the quantities of interest and the underlying assumptions. The model definition follows, detailing architectural choices, trade-off, and the implementation of the physics-informed loss. Finally, the performance of the proposed approach is discussed, including a comparison with traditional methods, and concluding remarks are presented. 2 Cislunar Orbit Determination 2.1 Scenario The mission analysis for the Gateway platform considers a Near-Rectilinear Halo Orbit (NRHO) with a period of approximately 7 days, located around the second collinear Euler–Lagrange point L_{2} of the Earth–Moon Circular Restricted Three-Body Problem (CR3BP), characterised by the mass parameter \mu=0.01215 . The Equations of Motion (EOM) written below are expressed in nondimensional units and in the synodic (rotating) reference frame Oxyz , centred at the barycentre of the Earth-Moon system, with x -axis pointing the Moon from the Earth, its z -axis in the angular momentum direction, and its y -axis completing the right-hand set [ 12 ] : \displaystyle\ddot{x} \displaystyle=2\dot{y}+U_{x}, (1a) \displaystyle\ddot{y} \displaystyle=-2\dot{x}+U_{y}, (1b) \displaystyle\ddot{z} \displaystyle=U_{z}. (1c) The state vector \mathbf{x}=[x,y,z,\dot{x},\dot{y},\dot{z}]^{T}\in\mathbb{R}^{6} collects, in order, the position and velocity components of the massless particle relative to the barycentre of the system in the non-inertial synodic frame. The quantities U_{x},U_{y},U_{z} denote the components of the gradient of the pseudo-potential function U(x,y,z) with respect to the position coordinates: U(x,y,z)=\frac{1}{2}(x^{2}+y^{2})+\frac{1-\mu}{r_{1}}+\frac{\mu}{r_{2}}+\frac{1}{2}\mu(1-\mu). (2) The scalar distances between the particle and the two primary bodies are respectively r_{1} and r_{2} : r_{1}=\sqrt{(x+\mu)^{2}+y^{2}+z^{2}},\ r_{2}=\sqrt{(x-(1-\mu))^{2}+y^{2}+z^{2}}. Figures 1 present a selection of ten trajectories belonging to the L_{2} -NRHO family. These orbits are obtained by numerically integrating the EOM ( 1 ) over one orbital period, which varies within approximately \pm 1 hour around the nominal 7-day period. (a) Synodic reference frame. The red stars denote the Euler-Lagrange points L_{1},L_{2} and the blue star the Moon (b) Geocentric inertial reference frame. The red line denotes the Moon’s orbit over 1 sidereal month Figure 1 : Selection of 10 orbits belonging to the L_{2} -NRHO family,