The appeal of so-called frozen orbits is well established. These are orbits whose size, shape, and orientation remain nearly constant in a frame rotating with the Moon, requiring almost no propellant to maintain. The Lunar Reconnaissance Orbiter demonstrated the payoff: by carefully selecting a polar frozen orbit, the mission used only 33.9 kilograms of hydrazine over twelve years for all of its station-keeping, phasing, and momentum management. Frozen orbits could similarly support extended surface observation campaigns, communications relays, and even the tracking of orbital debris or derelict hardware from past missions, such as the Apollo-era lunar module ascent stages that may still circle the Moon.

Traditionally, frozen orbits have been found analytically. Researchers construct an averaged dynamical model that strips out short-period oscillations, then solve for equilibrium points in the averaged equations. These classical analyses predict that frozen orbits occur under specific conditions: when the mean argument of perilune, the point of closest approach, sits near plus or minus 90 degrees; when the mean eccentricity is zero; or when the argument of perilune equals 0 or 180 degrees. The trouble is that averaging discards real physics, and the resulting predictions must be verified by propagating trajectories in a high-fidelity model that includes a detailed lunar gravity field and the true ephemerides of the Earth and Sun. Scaling that verification across the whole phase space has always been the bottleneck.

Bosanac and Miceli attacked the bottleneck directly. They defined over 1.1 million initial conditions at a fixed epoch, sampling orbital elements coarsely across semi-major axes up to 2,038 kilometers, eccentricities, inclinations from 0 to 180 degrees, and the full range of nodal longitudes and arguments of perilune. Each initial state was propagated for up to 180 days in a demanding dynamical model: a 100 by 100 degree-and-order spherical harmonic lunar gravity field, drawn from the GRGM900C model derived from GRAIL data, combined with the point-mass gravity of the Earth and Sun positioned according to the DE421 ephemerides. Integrations used a high-order Runge-Kutta scheme with tight tolerances of ten to the minus eleventh, terminating early if the spacecraft struck a spherical Moon of radius 1,738 kilometers.

Of the initial conditions tested, 638,181 trajectories survived at least three lunar rotational periods, roughly 81.9 days, and entered the analysis. Rather than storing every point along every path, the researchers represented each trajectory by its sequence of perilunes, the moments of closest approach, which capture the long-period evolution of the orbit’s eccentricity and orientation. Each perilune was described by two compact coordinates, p and q, equal to the eccentricity times the cosine and sine of the argument of perilune. These coordinates encode both the shape of the orbit and the direction in which its low point points, and they trace out distinctive loops and curves as the orbit evolves in the Moon-fixed frame.

The clustering itself relied on DBSCAN, a density-based algorithm that groups points lying in sufficiently crowded neighborhoods while labeling isolated points as noise. Because the dataset was too large to process at once, it was split into 64 partitions of roughly 10,000 trajectories each. Within each partition, trajectories were grouped if their perilune sequences moved through the pq-plane along similar paths at similar times, judged by unit tangent vectors sampled every half day. Local clusters were then merged across partitions and across different initial phasings, using shape-based feature vectors and a graph structure whose connected components defined the final groupings. The outcome was striking: 638,181 trajectories collapsed into just 885 global clusters, a reduction of three orders of magnitude in the information an analyst must examine.

From selected clusters, the team manually extracted sets of long-lifetime orbits, defined as trajectories showing limited drift in the pq-plane over the full 180 days. These orbits appear at discrete combinations of initial semi-major axis and inclination, and in many cases, shifting the initial longitude of the ascending node rotates the entire groundtrack geometry without changing the underlying orbital evolution. At a semi-major axis of 1,808 kilometers and an inclination of 85 degrees, for example, three different nodal longitudes produced long-lifetime orbits whose perilune groundtracks circled the north pole, circled the south pole, or swept across both hemispheres, all while the eccentricity and argument of perilune evolved in nearly identical fashion. That flexibility matters for mission design, because the groundtrack determines which terrain a spacecraft observes.

The comparison with classical theory was largely reassuring, with intriguing exceptions. Most of the recovered long-lifetime orbits cluster around arguments of perilune near plus or minus 90 degrees or encircle zero eccentricity in the pq-plane, exactly as analytical frozen orbit theories predict. The overall shapes of the mean eccentricity curves as a function of inclination also match the families derived by Lara in 2011 and later refined by Yarndley and colleagues. But at initial inclinations of 65 and 115 degrees, the numerical search surfaced low-drift orbits with mean arguments of perilune as far as 30 to 60 degrees away from the predicted minus 90 degrees. When the authors extended two of these trajectories to nearly twenty years, some remained tightly bounded, hinting at frozen orbits that the averaged theories miss, while others drifted substantially under longer-period oscillations governed by multiple frequencies.

The method also recovered orbits that analytical approaches cannot touch by construction. Resonant trajectories, whose perilune sequences form looping paths in the pq-plane, are typically excluded from averaged derivations, yet the clustering search found them naturally at low inclinations. Several results echoed earlier numerical work: orbits at 95 degrees inclination showed oscillation periods of 27.43 to 27.54 days, close to the 27.35-day repeat groundtrack period reported by Russell and Lara in 2007, and near-equatorial long-lifetime orbits resembled those designed for the DARE radio astronomy concept by Plice and colleagues.

The authors are careful to frame the work as a proof of concept. The initial conditions were sampled coarsely, at a single epoch, and only over 180 days, so some families may be missing and some orbits may not persist indefinitely. Extracting representative orbits from clusters is still manual, and automating that step is ongoing work. Even so, the demonstration is compelling: a data-mining pipeline can compress a million brute-force simulations into a compact, interpretable summary that agrees with decades of theory while exposing solutions theory overlooked. As NASA and its partners plan sustained lunar operations under Artemis, tools like this could quietly become as important as the rockets, ensuring that the spacecraft placed around the Moon stay there without burning fuel they cannot spare.

Subject of Research: Numerical identification of long-lifetime low lunar orbits using trajectory clustering in a high-fidelity lunar gravity model

Article Title: Numerical exploration of low lunar, long-lifetime orbits

Article References: Bosanac, N., & Miceli, G. E. (2026). Numerical exploration of low lunar, long-lifetime orbits. Celestial Mechanics and Dynamical Astronomy, 138(5), Article 61. https://doi.org/10.1007/s10569-026-10334-x

Image Credits: AI Generated

DOI: 10.1007/s10569-026-10334-x

Keywords: lunar orbits, frozen orbits, trajectory clustering, celestial mechanics, lunar gravity field, mission design, DBSCAN, perilune, space situational awareness, astrodynamics, Lunar Reconnaissance Orbiter, ephemeris model

The work, carried out by Jean Paulo dos Santos Carvalho of the Federal University of Recôncavo da Bahia in Brazil, revisits the classic critical inclination problem using the so-called extended Brown Hamiltonian model developed by Hantao Lei and Elke Grishin. Their formulation incorporates nonlinear corrections to the quadrupole-order disturbing potential that classical double-averaged theories ignore. These corrections turn out to be decisive in triple systems that are only weakly hierarchical — configurations where the perturbing body is not overwhelmingly distant, and where the standard approximations begin to break down. In such systems, Carvalho shows, the classical formula cos²i = 3/5, which fixes the critical inclination at 39.23 degrees for circular orbits, no longer tells the whole story.

The mathematics behind the classic result is elegant and deceptively simple. By averaging the third-body potential over both the orbit of the satellite and the orbit of the perturber, the long-period dynamics reduce to a single conserved quantity, the Kozai integral. Setting that integral at the separatrix value and then taking the limit of zero eccentricity yields the famous pair of critical inclinations: 39.23 degrees for prograde orbits and 140.77 degrees for retrograde ones. Below these thresholds, the argument of pericenter circulates freely through all 360 degrees; above them, it librates around equilibrium values while eccentricity and inclination exchange energy in large oscillations. For artificial satellite designers, that transition defines where orbital stability qualitatively changes.

Carvalho’s approach replaces the classical disturbing potential with the extended Brown Hamiltonian, whose extra terms are weighted by two small parameters, ε21 and ε22, that depend on the mass of the disturbing body, the masses of the central body and perturber, the semi-major axis and eccentricity of the satellite’s orbit, and the eccentricity and mean motion of the perturbing body’s apparent orbit. Substituting the modified potential into Lagrange’s planetary equations and demanding that the rate of change of the argument of pericenter vanish produces a new expression for the critical inclination — one that no longer depends on eccentricity alone but on the full physical configuration of the triple system.

The consequences are striking in a concrete application. For a spacecraft in a high-altitude orbit with a semi-major axis of 13,000 kilometers around the Moon, perturbed by Earth’s gravity, the classical theory predicts critical inclinations of 39.23 and 140.77 degrees at zero eccentricity. The modified model instead gives 42.65 and 144.13 degrees — a shift of more than three degrees in both the prograde and retrograde cases. At an eccentricity of 0.6, the same comparison moves the critical inclination from 51.71 degrees to 53.32 degrees. The greater the semi-major axis of the satellite’s orbit, the larger the deviation from the classical prediction, which is precisely what one would expect as the system becomes less hierarchical and higher-order effects grow stronger.

Beyond shifting the threshold, the extended model breaks a symmetry that has been built into the theory since Kozai’s original work. In the classical picture, the prograde and retrograde critical inclinations sit perfectly symmetric about 90 degrees: subtract either value from 90 degrees and the result is the same 50.77 degrees. With the Brown corrections included, that mirror symmetry vanishes — 90 minus 42.65 gives 47.35 degrees, while 144.13 minus 90 gives 54.13 degrees. This asymmetry, consistent with earlier analytical findings by Lei and Grishin, means that prograde and retrograde lunar orbits do not experience the Lidov–Kozai transition at mirror-image inclinations, a subtlety with real consequences for mission designers choosing orbital geometries around the Moon.

The study also maps the dynamics using e sin i versus e cos i diagrams, which reveal at a glance whether an orbit circulates, librates, or sits at a bifurcation. Under the classical model, the inclination of 39.23 degrees — the zero-eccentricity critical value — appears as a bifurcation curve even for an eccentric orbit at e = 0.6. Under the modified model, that special behavior disappears: 39.23 degrees falls below the new critical inclination of 53.32 degrees, and the corresponding orbit simply circulates. The classical number, in other words, is not merely imprecise in the modified framework — it loses its dynamical meaning for non-circular orbits, a fact Carvalho notes has been widely overlooked, since the 39.23-degree value is routinely quoted in the literature without reference to the eccentricity it was derived for.

To test the practical relevance of these results, the study examines a lunar analogue of the Molniya orbit: a frozen, highly eccentric, high-altitude trajectory around the Moon placed exactly at the new critical inclination of 53.32 degrees, with an eccentricity of 0.6 and an argument of perilune of 270 degrees. In the double-averaged modified model, the eccentricity remains constant, satisfying the frozen-orbit condition. When the same initial conditions are propagated with NASA’s open-source General Mission Analysis Tool, which models the third-body perturbation in full without averaging, the resulting orbit remains close to its initial elements — a quasi-frozen orbit. The classical Kozai model, by contrast, fails to reproduce the real trajectory, underlining how much accuracy is gained by adopting the extended potential.

One further finding carries operational weight for future lunar missions. In the single-averaged model, where only the satellite’s mean anomaly is eliminated, the longitude of the ascending node enters the equations explicitly. Carvalho shows that different initial values of this angle produce noticeably different orbital evolutions: with the node set to 180 degrees, the orbit exhibits smaller variations in its elements than when the node is set to 0 degrees, and simulations with NASA’s software confirm the same behavior. For planners of high-altitude, eccentric lunar orbits — including potential sites for lunar space stations — the initial orientation of the orbital plane is therefore not a detail but a design variable of primary importance.

The results vindicate the push behind the extended Brown Hamiltonian, which was originally validated against N-body simulations of Jupiter’s irregular satellites. Carvalho’s formula for the critical inclination, derived from the same potential but through a different route than Lei and Grishin’s perturbative approach, agrees with their equation to within a hundredth of a degree for the lunar case — a satisfying cross-check that two independent methods land on the same shifted thresholds. As space agencies prepare to place long-lived infrastructure in high lunar orbits, where Earth’s perturbation dominates the dynamics, the message of this study is clear: the safe, stable inclinations for such missions are not the textbook values inherited from 1962, but numbers that must be computed case by case from the masses, distances, and orbital elements of the bodies involved.

Subject of Research: The shift of the third-body critical inclination in celestial mechanics when nonlinear corrections to the classical Lidov–Kozai disturbing potential are included

Article Title: The problem of the critical inclination of the third body

Article References: dos Santos Carvalho, J. P. (2026). The problem of the critical inclination of the third body. Celestial Mechanics and Dynamical Astronomy, 138(5), Article 53. https://doi.org/10.1007/s10569-026-10328-9

Image Credits: AI Generated

DOI: 10.1007/s10569-026-10328-9

Keywords: critical inclination, Lidov-Kozai mechanism, celestial mechanics, three-body problem, lunar orbits, frozen orbits, extended Brown Hamiltonian, double-averaged model, third-body perturbation, orbital stability, Molniya orbit, space mission design

The mathematical heart of the new system is the Hori-Deprit method, a canonical perturbation technique based on Lie transformations that was developed in the 1960s by Gen-ichiro Hori and André Deprit. The method works on a Hamiltonian, the energy function that governs orbital motion, expanded in powers of a small parameter representing the strength of a perturbing force. Through a sequence of recursive canonical transformations, the algorithm systematically strips away the short periodic oscillations in the satellite’s motion, leaving behind a simplified, averaged Hamiltonian that captures only the long-term, secular evolution of the orbit. The approach rose to fame when Deprit and Rom used it in 1970 to produce an automated third-order solution to the main problem of artificial satellite theory, eliminating the small-eccentricity singularities that had plagued earlier analytical solutions.

What the new framework adds is a fully automated, open pipeline for carrying out such calculations at modern scale. The symcelmech package represents Hamiltonians as Poisson series, multivariate Fourier series whose coefficients are themselves Laurent series in the orbital variables, and it manipulates them within a closed algebraic domain so that differentiation, averaging and integration never produce unmanageable new mathematical objects. A dedicated Poisson bracket engine, equipped with memoization that caches previously computed brackets and exploits antisymmetry to avoid redundant work, drives the recursive Lie transformation. Because high-order analytical theories are often defeated not by physics but by expression swell, the explosive growth of symbolic terms, the authors devoted considerable effort to controlling that growth, including targeted rational simplification, selective trigonometric expansion on a single chosen variable, and eccentricity truncation that caps the number of terms entering each stage of the recursion.

The automation extends to the most computationally delicate step: integration. Maxima’s general-purpose symbolic integrator relies on heuristic searches that can buckle when asked to integrate expressions containing thousands of trigonometric terms. The researchers therefore wrote specialized routines that exploit the structure of Poisson series. Averaging is performed through a divide-and-conquer routine that classifies each term as secular or periodic in linear time, while quadrature for the generating functions is carried out by mapping trigonometric terms into the complex exponential domain, where integration becomes simple division by the angular frequency. For perturbations naturally expressed in the true anomaly, such as third-body gravity and solar radiation pressure, the framework offers closed-form averaging routes based on Hansen coefficients or on a change of variable to the eccentric anomaly, preserving exact eccentricity dependence rather than relying on truncated series.

Once the algebra is complete, the analytical expressions are exported as optimized text files and parsed by CelestialMechanics.jl, which converts them into fast Julia functions ready for numerical evaluation. The Julia toolkit assembles full-fidelity force models, including zonal harmonics up to degree 18, tesseral and sectorial gravity terms, third-body perturbations and solar radiation pressure with shadow models, drawing ephemerides from NASA’s SPICE library and gravity fields from standard planetary data files. It supports three distinct propagation modes, Cartesian Cowell integration, integration of Hamilton’s equations in Delaunay variables, and integration of Lagrange’s planetary equations in classical orbital elements, all feeding a unified post-processing pipeline that converts states into comparable orbital elements and generates scientific visualizations. For long-duration runs, trajectory data are streamed to disk to avoid exhausting memory.

The authors validated the pipeline in two increasingly demanding applications. In the first, they used symcelmech to normalize the Hamiltonian of the main J2 problem, the dominant perturbation caused by Earth’s equatorial bulge, through two successive Lie transformations that removed both the fast orbital angle and the argument of perigee. The entire symbolic computation, including a second-order generating function with thousands of terms, ran in a fraction of a second on an ordinary laptop. Predictions of the secular precession rates of perigee and ascending node derived from the doubly averaged Hamiltonian were then checked against two independent numerical propagations of the full, non-averaged J2 dynamics, agreeing to within roughly a part in a thousand, exactly consistent with the expected size of the neglected third-order contributions.

The second application targeted frozen orbits around the Moon, special orbits whose eccentricity and argument of perilune remain constant over time, making them ideal for stable low-altitude observation missions. The team built a high-fidelity lunar gravity model including zonal harmonics from J2 through J9 together with the leading tesseral and sectorial coefficients, and normalized the Hamiltonian in the Moon’s rotating body-fixed frame. Solving the frozen orbit conditions across a grid of semi-major axes and inclinations revealed distinct families of equilibria: bands of low-eccentricity frozen orbits at intermediate inclinations, a polar family with eccentricities below 0.01, and narrow strips of higher-eccentricity solutions near the equator driven by the odd zonal harmonics. Phase portraits confirmed the expected libration structure, with stable equilibria surrounded by closed energy contours and separatrices dividing librating from circulating orbits.

A representative near-polar frozen orbit was then propagated for 800 days, roughly nine thousand lunar revolutions, using an eighth-order symplectic integrator with a fixed one-minute step. The argument of perilune librated around 270 degrees with an amplitude of about 70 degrees, confirming that the orbit remained trapped in the libration island predicted by the analytical theory, while the periapsis altitude stayed confined between 165 and 185 kilometers. Crucially, the Jacobi constant, an exact integral of motion in the rotating frame, was conserved to about five parts in one hundred million with no secular drift, demonstrating that the long-term evolution observed in the simulation reflected genuine physics rather than numerical artifacts. The authors also documented a known limitation honestly: near-circular frozen orbits expose a coordinate singularity in Delaunay variables that will require reformulating the transformation in Poincare variables, a development already partially implemented.

Beyond the immediate results, the significance of the work lies in its openness and its architecture. The ecosystem of analytical celestial mechanics has long been dominated by proprietary mission analysis suites whose internal algorithms cannot be audited or extended by the community, and by aging research codes tied to obsolete languages. By releasing both symcelmech and CelestialMechanics.jl as open-source projects, the researchers hope to democratize access to sophisticated perturbation methods and invite collaborative refinement. Planned extensions include the elimination of the parallax in polar-nodal variables to tame memory demands at third order, computation of Birkhoff normal forms for stability analysis, non-singular Poincare formulations for near-circular orbits, and integration with the SciML ecosystem to explore physics-informed neural networks and machine-learned surrogate models for long-term orbit prediction. For an era of rapidly multiplying lunar and planetary missions, the framework offers a rigorous, transparent and computationally efficient foundation for designing the orbits of the future.

Subject of Research: An open-source hybrid symbolic-numerical framework automating Hori-Deprit perturbation theories for artificial satellite dynamics and space mission design.

Article Title: A hybrid symbolic-numerical framework for artificial satellite theory and dynamics using Maxima and Julia

Article References: de Oliveira Paes, G., Berton, L., & de Moraes, R. V. (2026). A hybrid symbolic-numerical framework for artificial satellite theory and dynamics using Maxima and Julia. Celestial Mechanics and Dynamical Astronomy, 138(5), Article 56. https://doi.org/10.1007/s10569-026-10331-0

Image Credits: AI Generated

DOI: 10.1007/s10569-026-10331-0

Keywords: celestial mechanics, artificial satellite theory, Hori-Deprit method, Lie transformations, computer algebra, Maxima, Julia, Poisson series, frozen orbits, lunar orbits, Hamiltonian dynamics, orbital perturbation theory