Acta Astronautica 229 (2025) 333–344

Contents lists available at ScienceDirect

Acta Astronautica
journal homepage: www.elsevier.com/locate/actaastro

Research paper

Direct low-energy trajectories to Near-Earth Objects

E. Fantino a ,∗, R. Flores a , G. Donnarumma a,b , D. Canales c , K.C. Howell d
a
Department of Aerospace Engineering, Khalifa University of Science and Technology, P.O. Box 127788, Abu Dhabi, United Arab Emirates
b
Department of Industrial Engineering, University of Naples Federico II, Piazzale Tecchio 80, 80125, Naples, Italy
c Department of Aerospace Engineering, Embry–Riddle Aeronautical University, Daytona Beach, FL 32117, USA
d School of Aeronautics and Astronautics, Purdue University, West Lafayette, IN 47907, USA

ARTICLE INFO ABSTRACT

Keywords: Near-Earth Objects (NEOs) are asteroids, comets and meteoroids in heliocentric orbits with perihelion below
Circular restricted three-body problem 1.3 au. Similarly to the population of the Main Asteroid Belt, NEOs are primordial bodies, and their study can
Hyperbolic invariant manifolds improve our understanding of the origins of the Solar System. With a catalog of over 30 000 known asteroids
Two-body problem
and approximately 100 listed short-period comets, the NEO population represents an inventory of exploration
Near-Earth Objects
targets reachable at a significantly lower cost than the objects of the Main Asteroid Belt. In addition, the
Rendezvous
Impulsive maneuvers
materials present in these bodies could be used to resupply spacecraft en route to other destinations. The
trajectories of past missions to NEOs have been designed with the patched-conics technique supplemented
by impulsive and/or low-thrust maneuvers and planetary gravity assist. The transfer times range from some
months to a few years, and the close-approach speeds relative to the target have been as high as 10 km/s. The
design technique described in this work leverages the invariant structures of the circular restricted three-body
problem (CR3BP) to connect the Earth’s vicinity with NEOs in low-eccentricity and low-inclination trajectories
in close proximity to the terrestrial orbit. The fundamental building blocks of the method are periodic orbits
around the collinear points L1 and L2 of the Sun-Earth CR3BP. These orbits are used to generate paths that
follow the associated hyperbolic invariant manifolds, exit the sphere of influence of the Earth and reach NEOs
on nearby orbits, thus enabling robotic as well as crewed exploration missions to targets in the terrestrial
region and asteroid deflection operations. The strategy is simple, can be applied to depart either a libration
point orbit or a geocentric orbit, and offers attractive performance features.

1. Introduction in 2010 [8]. Also in 2010, Deep Impact’s mission extension (EPOXI)
carried out a flyby with comet Hartley 2 [9]. In 2012, CNSA’s Change’e
Near-Earth Objects (NEOs) are asteroids (NEAs) and comets (NECs) 2 executed a close approach with 4179 Toutatis [10]. Between 2018
with perihelion below 1.3 au [1]. Aphelia of NEAs generally lie within and 2020, JAXA’s follow-up mission Hayabusa 2 rendezvoused with
a sphere of radius 5.2 au, defined by Jupiter’s orbit. Only short- asteroid Ryugu, collected samples of its surface and returned them
period comets (i.e., orbital period less than 200 years) are considered to Earth [11]. The S/C is now on an extended mission to asteroid
NECs [2]. Scientists believe that the characterization of these objects is 1998 KY26. NASA’s OSIRIS-Rex successfully rendezvoused with Bennu,
key to deepening our understanding of the origin and evolution of the touched down on its surface in 2020 and extracted samples that were
Solar System, and the source of water on Earth. These bodies have also delivered to Earth in 2023 [12,13]. The probe is continuing its journey
received attention because of their proximity to Earth, which is often to encounter and study 99 942 Apophis in 2029. The Double Asteroid
associated with collision threats [3,4]. Lastly, in recent years, NEOs Redirection Test (DART [14]) by NASA aimed at demonstrating a
have gained importance in the context of resource utilization, as in-situ method for planetary defense. In 2022, DART collided with Dimorphos,
collection and storage of material available in space could reduce the a satellite of the asteroid Didymos, and shortened its orbital period by
cost of space missions [5]. 32 min, proving the effectiveness of the transfer of momentum from
The first spacecraft (S/C) to visit a NEO was NASA’s NEAR Shoe- the S/C to the asteroid. The trajectories of these probes were designed
maker which landed on the surface of 433 Eros in 2001 [6]. This using a variety of techniques, including patched-conics, gravity assist,
endeavour was followed by NASA’s Deep Impact mission to comet impulsive and continuous-thrust maneuvers. The most relevant perfor-
Tempel 1 in 2005 [7]. The same year, JAXA’s Hayabusa probe col- mance figures of these missions are summarized in Table 1.
lected samples from asteroid Itokawa and delivered them to Earth

∗ Corresponding author.
E-mail address: [email protected] (E. Fantino).

https://doi.org/10.1016/j.actaastro.2025.01.020
Received 24 June 2024; Received in revised form 31 October 2024; Accepted 7 January 2025
Available online 14 January 2025
0094-5765/© 2025 The Authors. Published by Elsevier Ltd on behalf of IAA. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 1. Left: propagation of branches of hyperbolic invariant manifolds associated with PLOs around L1 (red) and L2 (blue) in the Sun-Earth rotating reference frame (the green
ring represents the forbidden region for the specific energy level considered). Right: orbits of several NEOs and the terrestrial planets. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)

Table 1
Trajectory information of past missions to NEOs: transfer technique, launch 𝐶3 , time of flight to destination (𝛥𝑡), 𝛥𝑉 budget of deterministic
maneuvers, target approach mode (outcome) and velocity relative to target at arrival (𝑉𝑟 ). Meaning of abbreviations and symbols: GA = gravity
assist; PC = patched conics; HT = high thrust; LT = low thrust; FB = flyby; SL = soft landing; I = impact; - = data not found.
Mission Technique Launch 𝐶3 𝛥𝑡 𝛥𝑉 Outcome 𝑉𝑟
(km2 /s2 ) (year) (m/s) (km/s)
NEAR [15] GA/PC 25.9 5.0 1176 SL 0.0
Deep Impact [7] HT 11.75 ∼0.5 – FB+I 10
EPOXI [9] GA/PC Extension 8.0 ∼150 FB+I 12.3
Hayabusa [8] LT – 2.5 – SL 0.0
Chang’e 2 [10] via Moon + L2 Lissajous 2.2 677 – FB 10.7
Hayabusa 2 [11] GA/LT 21.0 3.6 ∼2000 SL 0.0
OSIRIS-Rex [12] GA/HT 29.3 2.4 ∼900 SL 0.0
DART [16] HT/LT – 0.8 – FB+I 6.0

NEAs reachable with HT or LT, directly or via GAs, between 2020 and
2024 and accessible for rendezvous or sample collection and return.
Human exploration of NEAs has gained a lot of interest too. Since
2012, NASA has been updating the Near-Earth Object Human Space
Flight Accessible Target Study (NHATS) database of targets on orbits
very close to Earth’s and reachable by a round-trip mission of limited
duration [24]. The trajectory design is carried out with the method of
patched-conics, and employs full-precision ephemerides for both the
S/C and the asteroids. Recently, Sánchez & Yárnoz [25] proposed a
method to drive objects from accessible heliocentric orbits into the
Earth’s neighborhood using LT and the stable invariant manifolds of
Lyapunov and halo orbits. Jorba & Nicolás [26] focused on the cap-
ture of NEAs by means of approximations of the stable manifolds of
hyperbolic tori associated with the L3 point of the Earth-Moon system
in the planar Earth-Moon-Sun bicircular problem. These past efforts do
Fig. 2. The Sun-Earth synodic barycentric reference frame (𝑂𝑥𝑦, rotating frame) and
the barycentric ecliptic J2000.0 reference frame (𝑂𝑋 𝑌 , inertial frame). 𝑂 is the not address the systematic use of low-energy trajectories to construct
barycenter of the Sun-Earth system, 𝑆 the Sun and 𝐸 the Earth. direct transfers and rendezvous opportunities.
This work introduces a simple approach to identify a comprehensive
set of low-energy trajectories to reach NEOs. The method exploits the
The Asteroid Redirect Mission (ARM [17]) was a solar-electric dynamical and geometrical features of the hyperbolic invariant mani-
folds associated with periodic orbits around the collinear equilibrium
propulsion mission proposed by NASA to rendezvous with a large NEA
points L1 and L2 of the Sun-Earth circular restricted three-body problem
(2011 MD was a prime candidate), use robotic arms to retrieve a
(CR3BP). These trajectories extend into near-Earth space both inside
4-meter boulder, and bring it into lunar orbit. In fact, methods to ren-
and outside of the terrestrial orbit. They also acquire a phase differ-
dezvous, retrieve and deliver portions or entire asteroids to near-Earth ence with respect to Earth’s motion, spanning a wide range of orbital
space are a major focus of investigations in the context of missions to domains and coming in close proximity to many resident objects. This is
NEOs. The trajectory design approaches include conventional patched- evidenced by Fig. 1, depicting the paths of unstable hyperbolic invari-
conics, direct Lambert arcs and gravity assist (GA), and the propulsion ant manifolds associated with planar Lyapunov orbits (PLOs) around
can be chemical (high thrust, HT) or electrical (low-thrust, LT) (see, L1 (red) and L2 (blue) in the Sun-Earth rotating reference frame (left),
e.g., [18–22]). Notably, Strange et al. [23] provide an extensive list of and the orbits of several NEOs together with the terrestrial planets

334
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 3. Families of 50 PLOs around L2 (left) and L1 (right) in the synodic barycentric reference frame.

Table 2
Physical parameters of the Sun-Earth system used in this work [34].
Symbol Definition Value Units
𝐺 𝑀𝑆 Sun’s gravitational parameter 1.3271244 ⋅ 1011 km3 /s2
𝐺 𝑀𝐸 Earth’s gravitational parameter 3.9860044 ⋅ 105 km3 /s2
𝑅𝐸 Earth’s equatorial radius 6378.1366 km
𝜇 Earth-Sun mass ratio 0.30034806 ⋅ 10−5 –
𝑑 Earth-Sun distance 149 597 870 km

The application of the hybrid CR3BP/2BP to the design of trajecto-

ries to NEOs is novel. This approximate model is one step above simple
patched-conics in fidelity. Therefore, it is adequate for preliminary
analyses. For detailed studies, it can be extended to an ephemeris
model including the gravity of additional bodies and non-gravitational
perturbations.
The paper starts with the definition of the dynamical model for
the S/C (Section 2). Section 3 deals with the selection of the tar-
get NEOs and the processing of their orbital data. Sections 4 and
Fig. 4. Generation of initial conditions for TOs inside a PLO with 𝐶𝐽 = 3.00057: initial
70 × 70 grid (gray), internal points (orange), accepted internal points that connect with
5 present the methodology and results for planar and 3D transfers,
a 300-km altitude circular orbit around the Earth (red). For the sake of clarity, the two respectively. Section 6 compares their performance against conven-
axes have been drawn with different scales. (For interpretation of the references to tional patched-conics trajectories. Section 7 summarizes and concludes
color in this figure legend, the reader is referred to the web version of this article.) the paper. Preliminary explorations of the method can be found in
Donnarumma [31], and have been presented at the 2nd International
Stardust Conference [32] and at the 74th International Astronautical
(right). These natural pathways are characterized by low energies, Congress [33].
i.e., low velocities with respect to Earth, such as those typical of CR3BP
trajectories, which makes them accessible at very low launch cost. A 2. Dynamical model
key feature of the method is the approximation of these three-body
trajectories with heliocentric two-body (2BP) ellipses beyond a certain The motion of the S/C starts in the vicinity of the Earth and is
distance from Earth. This technique, also called patched CR3BP/2BP governed by the gravitational attraction of the Sun (mass 𝑀𝑆 ) and
model, was developed to design low-energy transfers between Galilean the Earth (mass 𝑀𝐸 ), i.e., the dynamical model is the Sun-Earth-S/C
moons, simplifying the computation of connections between invariant CR3BP. If the gravitational constant 𝐺, the sum 𝑀𝑆 + 𝑀𝐸 and the
manifold trajectories departing and approaching pairs of moons [27– distance 𝑑 between the primaries are used as reference magnitudes, the
30]. In a similar way, in this work the encounter between the S/C and mean motion of the primaries becomes unitary and their orbital period
a NEO occurs at the intersection between the heliocentric osculating equals 2𝜋. In the synodic barycentric reference frame, the Sun and the
Keplerian orbits of the two bodies, and the rendezvous becomes a sim- Earth are located at (𝜇,0,0) and (𝜇 − 1,0,0), where 𝜇 denotes the mass
ple analytical problem. An impulsive maneuver at an intersection point ratio 𝑀𝐸 ∕(𝑀𝑆 +𝑀𝐸 ) of the system (see Fig. 2). Table 2 lists the relevant
yields a zero relative velocity encounter between the S/C and the target. physical parameters of the Sun-Earth system.
Since the initial conditions of the trajectory of the S/C are generated in Fig. 3 illustrates families of PLOs around the two collinear equilib-
the Sun-Earth synodic barycentric reference frame, the orientation in rium points L1 and L2 . Each set contains 50 orbits equally spaced in
space of the corresponding heliocentric ellipse varies linearly with the Jacobi constant 𝐶𝐽 between 3.00056 (𝑦 amplitude ≃ 3 ⋅ 106 km) and
departure epoch. This feature provides a degree of freedom enabling 3.00089 (𝑦 amplitude ≃ 9 ⋅ 103 km).
the selection of the most desirable solutions, for example on the basis The PLOs serve to generate two types of trajectories:
of the 𝛥𝑉 budget or the transfer time 𝛥𝑡. The methodology was de-
veloped in the planar approximation for low-inclination targets, and • Outward branches of hyperbolic invariant manifolds (MTs) (200
subsequently extended to 3D solutions. trajectories for each PLO). They are determined and propagated

335
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 5. Characteristic launch energies of the TOs associated with L1 (left) and L2 (right).

The accepted TOs and all the MTs are propagated forward in time
until they intersect an Earth-centered circle, called circle of influence
(CI, Fig. 7) whose radius 𝑟𝐶 𝐼 equals 2 773 940 km or three times
𝑟𝑆 𝑜𝐼 = 𝑑 (𝑀𝐸 ∕𝑀𝑆 )2∕5 , the radius of the Laplace sphere of influence
of the Earth. This choice ensures that the CI encloses all the PLOs of
the two families, as shown in Fig. 8.
MTs and TOs represent different mission opportunities. MTs can
be used, for example, to depart the vicinity of an equilibrium point,
whereas TOs are direct trajectories from a geocentric parking orbit. The
two types span the same energy levels, but while MTs initially stay close
to the LPO, TOs pass through it and reach the CI directly, resulting in
faster transfers (see Fig. 9).
The state vectors of MTs and TOs at the CI are collected and
transformed to the heliocentric ecliptic J2000.0 reference frame. In
this operation, the distance between the barycenter of the Sun-Earth
system and the center of the Sun is neglected. The longitude of the
perihelion of the osculating heliocentric ellipses depends on the date
Fig. 6. Number of TOs that reach a 300 km Earth’s parking orbit per equilibrium point
of CI crossing, and varies at a constant rate equal to the mean motion
as a function of 𝐶𝐽 .
𝑛𝐸 𝑆 of the Sun-Earth system. Semimajor axis, eccentricity and true
anomaly depend entirely on the state vector at the CI in the synodic
frame. This is sketched in Fig. 10 which shows two osculating ellipses
using standard methods. An initial state is generated by applying corresponding to the same three-body state vector at two different CI
a small perturbation in the outward direction of the unstable crossing times, 𝑡1 and 𝑡2 (resulting from two different departure dates
eigenvector of the monodromy matrix of the PLO after appropri- on the same CR3BP trajectory). The two orbits have the same shape and
ate time transformation through the state transition matrix (see have the same true anomaly at CI crossing. The directions of the lines
e.g., [35]), followed by globalization of the manifold by forward of apsides are related through 𝜔𝑆∕𝐶 (𝑡2 ) = 𝜔𝑆∕𝐶 (𝑡1 ) + 𝑛𝐸 𝑆 (𝑡2 − 𝑡1 ), where
time propagation; 𝜔𝑆∕𝐶 is the longitude of the perihelion of the S/C’s orbit. The value of
• Transit orbits (TOs), obtained from 70 × 70 equally-spaced points this parameter depends on the angle (𝛼 in Fig. 2) between the synodic
of a rectangular grid bounding the PLO (Fig. 4). Each point frame and the inertial frame at the time of CI crossing. This angle is
internal to the PLO (orange markers) is the start of a TO with zero at the Spring Equinox, whose reference epoch is 2 460 024.3917
the same 𝐶𝐽 as the PLO. The direction of the initial velocity JD = 2023-Mar-20 21:24:00 UTC. This date is used to determine the
is parallel (L1 ) or anti-parallel (L2 ) to the 𝑥-axis.1 In this way, instantaneous orientation of the 𝑥-axis of the synodic frame.
the TOs drive the S/C away from the Earth forward in time
3. NEOs: orbital data and dynamical model
along paths contained in the outward invariant manifold branch
with the same 𝐶𝐽 . The initial states of the TOs are propagated
The orbital parameters of the candidate targets for rendezvous
backwards in time (i.e., towards the Earth). The trajectories that
have been extracted from the Solar System Dynamics Minor Bodies
cross a circular parking orbit at 300 km altitude are retained (red
Database [36], which on July 30, 2023 listed orbital data for 32 396
markers in Fig. 4), the others are discarded. Fig. 5 shows the asteroids and comets with perihelion below 1.3 au. Retaining only
characteristic launch energies (𝐶3 ) of the retained TOs, computed objects with ecliptic inclination ≤5◦ , perihelion distance ≥0.9 au and
in the Earth-S/C 2BP. The number of retained TOs varies with aphelion distance ≤1.1 au yields a set of 72 asteroids (Table 3). The
the amplitude (hence, the value of 𝐶𝐽 ) of the PLO. Only the first semimajor axes range from 0.971 to 1.058 au, while the eccentricities
20 Lyapunov orbits of each family provide solutions that reach span the interval [0.039, 0.0898]. Fig. 11 shows the distribution of
Earth’s vicinity (Fig. 6). orbital inclinations and the intervals of heliocentric distances swept by
the 72 orbits. All data are given in the heliocentric ecliptic J2000.0 ref-
erence frame with osculation epoch 𝑡0 = 2 460 000.5 JD = 2023-Feb-25
1
Other choices are possible because the position and the Jacobi constant 00:00:00 UTC. The dynamical model for the NEOs is the Sun-asteroid
only constrain the magnitude of the velocity. 2BP.

336
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 7. Left: propagation of TOs from the accepted grid points towards the parking orbit (backward) and the CI (forward). Right: forward propagation of the outward branches
of two MTs until they cross the CI.

Table 3
Candidate targets: primary designation and object index assigned in this work.
1 = 459872 16 = 2012 LA 31 = 2017 FT102 46 = 2020 HO5 61 = 2022 NX1
2 = 1991 VG 17 = 2012 TF79 32 = 2017 HU49 47 = 2020 MU1 62 = 2022 OB5
3 = 2000 SG344 18 = 2013 BS45 33 = 2018 FM3 48 = 2020 RB4 63 = 2022 RS1
4 = 2003 YN107 19 = 2013 GH66 34 = 2018 PK21 49 = 2020 VN1 64 = 2022 RD2
5 = 2006 JY26 20 = 2013 RZ53 35 = 2018 PN22 50 = 2020 WY 65 = 2022 RW3
6 = 2006 QQ56 21 = 2014 DJ80 36 = 2018 PM28 51 = 2021 AT2 66 = 2023 FY3
7 = 2006 RH120 22 = 2014 QD364 37 = 2018 WV1 52 = 2021 AK5 67 = 2023 GQ1
8 = 2008 KT 23 = 2014 WU200 38 = 2019 FV2 53 = 2021 CZ4 68 = 2023 GT1
9 = 2008 UA202 24 = 2014 WX202 39 = 2019 GF1 54 = 2021 GM1 69 = 2023 HM4
10 = 2010 JW34 25 = 2015 JD3 40 = 2019 KJ2 55 = 2021 LF6 70 = 2023 HG11
11 = 2010 VQ98 26 = 2015 XZ378 41 = 2019 PO1 56 = 2021 RZ3 71 = 2023 LE
12 = 2011 BL45 27 = 2016 GK135 42 = 2020 CD3 57 = 2021 RG12 72 = 2023 LG2
13 = 2011 MD 28 = 2016 RD34 43 = 2020 FA1 58 = 2021 VH2
14 = 2011 UD21 29 = 2016 YR 44 = 2020 GE 59 = 2021 VX22
15 = 2012 FC71 30 = 2017 BN93 45 = 2020 HF4 60 = 2022 BY39

plane. Hence, the rendezvous occurs at the intersection between their

coplanar confocal elliptical orbits. The solution of this mathematical
problem is due to [37] and was recently applied to design transfers
between giant planet moons [27–30]. The semimajor axis and the
eccentricity of the orbits of the two bodies, and the angle between the
two periapses determine the number of intersection points. As shown
in Fig. 12, which portrays two elliptical orbits with their arguments
of perihelion 𝜔𝑁 𝐸 𝑂 and 𝜔𝑆∕𝐶 , the number of intersection points can
be 0, 1 (tangential intersections A = B) or 2 (A ≠ B) depending on
the relative sizes and orientations of the ellipses. The tangential case
corresponds to the rendezvous impulse 𝛥𝐕# = 𝐕𝑆∕𝐶 − 𝐕𝑁 𝐸 𝑂 with the
smallest magnitude (see, e.g., [27]).
The longitude of the orbit perihelion for the NEO is fixed at the
value corresponding to the osculation epoch. For the S/C this quantity
is a function of the CI crossing date, which ultimately depends on
the departure date. The latter is computed a posteriori on the basis
of the rendezvous requirements, as shown later.2 The range of values
of the longitude of the perihelion of the orbit of the S/C for which
Fig. 8. The two families of PLOs and the CI. intersections with the target orbit exist can be determined analytically
(see [27]). This range is sampled with a uniform step 𝛥𝜔, and each
value (denoted 𝜔𝑑 meaning desired 𝜔) is used to compute the position
of the intersection point(s) between the elliptical orbits. The time of
4. Planar impulsive rendezvous
passage of the target NEO through the rendezvous point determines
the rendezvous date 𝑡# (Fig. 13 left). From this epoch, the (desired)
Neglecting the gravitational attraction of the Earth outside the CI
transforms the design of a direct transfer to a NEO into a search for
intersections between confocal ellipses. In the planar approximation, 2
The autonomous character of the CR3BP ensures freedom in the choice
the selected low-inclination targets and the S/C move in the ecliptic of the departure date.

337
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 9. Transfer times from each departure location to the CI for each trajectory type.

asteroid at 𝑡# , leading to a two-impulse transfer, i.e., 𝛥𝐕𝐶 𝐼 at the CI and

𝛥𝐕′# at rendezvous (Fig. 13 right). Alternatively, 𝛥𝑟# can be reduced or
eliminated by replacing the impulsive rendezvous maneuver (𝛥𝐕# ) with
a LT transfer.
The strategy outlined above has been applied to all the combina-
tions of target orbits and S/C trajectories (MTs and TOs from L1 and L2 ),
delivering a comprehensive coverage of the solution space that benefits
from the synergy between the CR3BP and 2BP approximations. The
simulations have been conducted with sampling intervals 𝛥𝜔 of 0.4◦
and 𝜔𝑒 = 0.02◦ . Setting an upper limit of 1500 m/s on the magnitude
Fig. 10. Effect of the CI crossing time (𝑡1 < 𝑡2 ) on the longitude of the perihelion 𝜔𝑆∕𝐶
of 𝛥𝐕# yields close to 70 000 rendezvous trajectories to a subset of
of an osculating heliocentric orbit obtained from the same CR3BP state vector at the 24 asteroids from the initial list. Table 4 lists the number of solutions
CI. and the features of the transfers with the best performance for each
target and trajectory type: i.e., minimum 𝛥𝑉# (𝛥𝑉#𝑚 ) and corresponding
transfer time, minimum 𝛥𝑡 (𝛥𝑡𝑚 ) and corresponding rendezvous im-
time 𝑡𝐶 𝐼 of CI crossing is computed. This date is associated with a pulse, and minimum 𝛥𝑟# (𝛥𝑟#𝑚 ). Fig. 14 illustrates six solutions from the
value (denoted 𝜔𝑟 or real 𝜔) of the longitude of the perihelion of the set. Table 5 compares the magnitude 𝛥𝑉# of the rendezvous maneuver
S/C orbit. A non-zero 𝜔𝑒 = ∣ 𝜔𝑑 − 𝜔𝑟 ∣ yields a non-zero rendezvous with the cost of the two-impulse transfer (𝛥𝑉𝐶 𝐼 , 𝛥𝑉#′ ) for the six cases
distance 𝛥𝑟# between the S/C and the target at 𝑡# (Fig. 13 middle). portrayed in Fig. 14.
𝛥𝑟# depends on 𝜔𝑒 and the eccentricity of the two orbits. In the limit
case of zero eccentricity and orbital radius = 𝑟 for both the S/C and 5. Extension to 3D impulsive rendezvous trajectories
the target, 𝛥𝑟# = 2𝑟 sin(𝜔𝑒 ∕2), which for 𝜔𝑒 = 0.01◦ and 𝑟 = 1 au
gives 𝛥𝑟# ≃ 104 km. Therefore, limiting the value of 𝜔𝑒 (by discarding Impulsive rendezvous trajectories taking into account the 3D ori-
solutions for which this quantity exceeds a pre-assigned threshold) entation (inclination, longitude of the ascending node, argument of
leads to encounters with an acceptable close-approach distance 𝛥𝑟# . perihelion) of the orbit of the candidate NEO have been designed by
Then, 𝛥𝑉# and the total transfer time 𝛥𝑡 can be used to characterize computing the intersections between confocal non-coplanar elliptical
the performance of a rendezvous trajectory with the given target. orbits, one with fixed orbital elements (NEO) and the other lying on the
Note that 𝛥𝑟# can be reduced to zero, for instance, by computing ecliptic plane and oriented as determined by the date of CI crossing.
the Lambert arc between the position of the S/C at 𝑡𝐶 𝐼 and that of the Under this assumption, the intersection must occur at either node of

338
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 11. Orbital inclinations (left) and perihelion - aphelion ranges (right) of the 72 selected NEOs.

Fig. 12. Number of intersection points between confocal coplanar ellipses for different mutual orientations: 0 (left), 1 (center), 2 (right).

Fig. 13. Left: Ideal S/C transfer and rendezvous with 𝜔𝑆∕𝐶 = 𝜔𝑑 . Middle: real S/C transfer with 𝜔𝑆∕𝐶 = 𝜔𝑟 . Right: Two-impulse transfer through the solution of Lambert’s problem
between 𝑡𝐶 𝐼 and 𝑡# .

the target’s orbit. However, the correct phasing is not guaranteed a 6. Comparison with patched-conics solutions
priori. The algorithm adjusts the CI crossing date (which determines the
departure date) and calculates a ballistic heliocentric arc connecting The performance of the 3D low-energy impulsive transfers via TOs
the CI with a node of the target’s orbit. A Lambert solver computes the described in this contribution have been compared with the trajectories
necessary impulses at the CI (𝛥𝐕1 ) and the rendezvous point (𝛥𝐕2 ) (see of the mission concept study carried out within NASA’s Planetary
Fig. 15). Science Decadal Survey devoted to NEAs [23]. It provides direct, HT
rendezvous trajectories with launch dates between 1-Jan-2020 and 1-
The simulations consider both types of trajectories (MTs and TOs)
Jan-2025, maximum flight time to the target of 5 years, launch C3
using a discretization of 5 days for the CI crossing date and setting
below 25 km2 /s2 , and rendezvous 𝛥𝑉 under 3 km/s. NASA’s solutions
bounds of 2.5 km/s for the total transfer impulse (𝛥𝑉 = 𝛥𝑉1 + 𝛥𝑉2 ) and are computed using a simplified, suboptimal, patched-conics approx-
1.2 years for the time of flight. The results are summarized in Tables 6 imation and then optimized with respect to the delivered S/C mass
and 7, which for each accessible target and trajectory type yields the at rendezvous via Sequential Quadratic Programming. The work of
number of solutions obtained, minimum 𝛥𝑉 (𝛥𝑉𝑚 ) and corresponding Strange et al. [23] was published in 2010, when the majority of the
transfer time, fastest transfer (𝛥𝑡𝑚 ) and corresponding 𝛥𝑉 . Fig. 16 asteroids considered in this work were yet to be discovered. As a result,
illustrates the ecliptic projection and the 3D view of two trajectories. the intersection of the two datasets contains only two targets: 2003

339
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Table 4
Characteristics of the best planar rendezvous trajectories to the selected NEAs. The solutions inside a box are illustrated in Fig. 14.
Ind. No. of solutions MT-L1 MT-L2 TO-L1 TO-L2
MT-L1 /MT-L2 𝛥𝑉#𝑚 (𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑟#𝑚 𝛥𝑉#𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑟#𝑚 𝛥𝑉#𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑟#𝑚 𝛥𝑉#𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑟#𝑚
/TO-L1 /TO-L2 m/s (day) day (m/s) km m/s (day) day (m/s) km m/s (day) day (m/s) km m/s (day) day (m/s) km
35 609/0/34/0 357 (392) 354 (491) 82 – – – 419 (300) 250 (543) 1 973 – – –
39 684/0/0/0 716 (554) 478 (990) 17 – – – – – – – – –
44 0/882/0/198 – – – 344 (402) 340 (535) 9 – – – 359 (254) 186 (549) 5
45 0/2981/0/0 – – – 378 (666) 552 (994) 92 – – – – – –
46 8042/0/5157/0 405 (643) 470 (818) 8 – – – 405 (419) 310 (997) 1 – – –
47 0/330/0/0 – – – 676 (675) 637 (922) 113 – – – – – –
49 1488/0/7/0 464 (468) 399 (921) 10 – – – 572 (334) 324 (811) 8 905 – – –
50 0/2060/0/0 – – – 558 (717) 540 (991) 134 – – – – – –
52 0/1737/0/0 – – – 394 (693) 604 (995) 3 – – – – – –
54 2065/0/34/0 332 (612) 531 (997) 12 – – – 500 (480) 463 (702) 11 848 – – –
55 1214/0/0/0 596 (564) 476 (996) 16 – – – – – – – – –
56 0/1025/0/10 – – – 585 (550) 484 (994) 1 – – – 593 (448) 395 (996) 2344
59 0/1251/0/43 – – – 492 (543) 447 (859) 161 – – – 499 (395) 356 (750) 128
61 0/3058/0/209 – – – 133 (530) 427 (357) 34 – – – 130 (378) 301 (271) 96
63 0/375/0/0 – – – 624 (582) 498 (996) 301 – – – – – –
64 0/1076/0/24 – – – 133 (513) 416 (240) 18 – – – 132 (359) 339 (174) 1329
65 0/116/0/49 – – – 152 (513) 510 (164) 305 – – – 168 (354) 330 (173) 4724
66 0/6233/0/814 – – – 432 (397) 277 (995) 2 – – – 447 (152) 98 (644) 5
67 0/8065/0/206 – – – 381 (402) 299 (814) 26 – – – 599 (268) 216 (939) 280
68 0/6183/0/5307 – – – 133 (407) 265 (585) 7 – – – 131 (160) 90 (566) 10
69 0/4767/0/65 – – – 462 (376) 295 (930) 1 – – – 694 (281) 236 (711) 129
70 0/2562/0/84 – – – 244 (381) 320 (542) 3 – – – 275 (280) 229 (500) 731
71 0/1560/0/49 – – – 548 (411) 323 (911) 59 – – – 590 (274) 221 (959) 19
72 15/0/0/0 979 (409) 398 (997) 455 – – – – – – – – –

Table 5 duration and lower total impulse). This demonstrates the ability of the
Magnitude of the rendezvous maneuver (second column) and cost of the two-impulse
new technique to find highly-efficient solutions, even when the time of
transfer (third and fourth column) for the six cases portrayed in Fig. 14.
flight is constrained. It is a valuable tool to expand the solution space
Case 𝛥𝑉# 𝛥𝑉𝐶 𝐼 𝛥𝑉#′
(m/s) (m/s) (m/s)
covered by more traditional methods.
a 332.1 0.46 332.6
b 995.5 20.5 992.8
7. Discussion and conclusions
c 693.6 1.11 694.6
d 405.7 0.15 405.5 This original contribution presents a simple and efficient technique
e 131.9 0.03 131.9 to systematically generate low-energy impulsive rendezvous trajec-
f 959.5 10.1 962.2
tories to Near-Earth Objects on low-eccentricity, low-inclination or-
bits in close proximity to the terrestrial orbit. The dynamical model
is the patched Sun-Earth circular restricted three-body problem/Sun-
YN107 (object index 4) and 2006 QQ56 (object index 6). Table 8 spacecraft two-body problem. Hence, the influence of the Sun is in-
compares total transfer 𝛥𝑉 , time of flight 𝛥𝑡 and launch C3 from corporated along the entire path. Two types of initial conditions are
Appendix C of [23] against the 3D TO/L1 transfers with the lowest 𝛥𝑉 considered: departures from planar Lyapunov orbits around Sun-Earth
(see Table 6) to the same targets. Note that the solutions by Strange L1 or L2 via outward branches of the associated hyperbolic invariant
et al. [23] launch approximately 2.5 years earlier than those obtained manifolds, and launches from a 300-km altitude circular orbit via
in this work. selected transit orbits (trajectories internal to the same hyperbolic
The cost of the rendezvous with 2003 YN107 is significantly higher invariant manifolds). Thus, two mission scenarios are accommodated:
when traveling via a TO (2321 m/s versus 1250 m/s), but this disad- direct Earth departures and transfers from an intermediate gateway
vantage is partially offset by a lower launch energy. Assuming a 300-km near a Lagrange point. The two types of trajectories exhibit different
altitude circular departure orbit, the two values of the launch energy performance characteristics, particularly in terms of time of flight.
(0.13 km2 /s2 and 6.30 km2 /s2 ) map into departure 𝛥𝑉 ’s of 3212 m/s Earth departures are generally faster due to their specific geometrical
and 3490 m/s. Thus, the net advantage of NASA’s solution is 793 m/s. characteristics.
Note, however, that this comes at the expense of a much longer transfer The heliocentric osculating two-body orbit obtained by transform-
duration (almost double). It is generally possible to reduce propellant ing the state of the spacecraft at the crossing of Earth’s sphere of
consumption by increasing the time of flight, so it is not surprising influence lies on the ecliptic plane. The mean motion of the orbit of the
that the longer transfer is more efficient. Note also that the search for Earth around the Sun along with the departure date determine the lon-
feasible trajectories limited the time of flight to 1.2 years. The limit gitude of the perihelion of such orbit, providing a fundamental degree
made it impossible to find a close match of NASA’s solution, but it is not of freedom (the optimization variable) of the method, which calculates
an intrinsic limitation of the methodology. Raising the maximum time geometrical intersections between the elliptical orbits of the spacecraft
of flight for the search enables revisiting the intersection points of the and the target (in the planar case, the latter is projected on the ecliptic).
heliocentric ellipses once per orbital period of the S/C. Thus, additional Thus, when the rendezvous conditions are met, the transfer requires a
phasing conditions can be analyzed for improved fuel-efficiency. This single impulsive maneuver at arrival, the cost of which depends on the
enables a comprehensive exploration of the solution space, making the choice of the intersection point used for the encounter. The extension of
technique quite versatile. the method to inclined target orbits introduces a Lambert arc between
The trajectory to 2006 QQ56 proposed in this work (see also case the sphere of influence of the Earth and one of the nodes of the orbit
h, Fig. 16) outperforms the solution by [23] in all aspects (shorter of the target, requiring two impulses.

340
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 14. Examples of planar rendezvous trajectories.

automatically emerges from the variation of the departure date as part

of the search for the rendezvous conditions. The simple dependence
of the orbital elements of the heliocentric orbit of the spacecraft on
the departure state and date helps reduce the computational burden.
Furthermore, the propagation of transit orbits and invariant manifold
trajectories to the boundary of the sphere of influence of the Earth
needs to be carried out only once. All these features make the method
computationally efficient. The entire simulation for 72 candidate tar-
gets using the full set of invariant manifolds and transit orbits through
both libration points took approximately 40 min in a mid-range laptop
when programmed in the Fortran language.
It is important to note that the method here described identifies all
the direct rendezvous opportunities that exist in the selected conditions
(resolution, tolerances and maximum allowed 𝛥𝑉 budget and transfer
time) at the chosen energy levels and in the approximation offered by
the dynamical model. Hence, an extension to include other types of
Fig. 15. Schematic representation of a two-impulse transfer to rendezvous with a target departure conditions would lead to a general tool capable of generating
in an inclined orbit. The encounter takes place at one of the nodes. optimal direct interplanetary trajectories.
The patched three-body/two-body approximation adopted is at least
as accurate as the conventional patched-conics approach. Hence, it is
In 2D, the identification of the rendezvous conditions is an analyti- suitable for feasibility analyses, and its results can be refined with
cal problem that can be solved in closed form. The 𝛥𝑉 -optimal transfer an ephemeris model for higher-fidelity mission design. The impulsive

341
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Table 6
Characteristics of the best 3D rendezvous trajectories to the selected NEAs with 𝛥𝑉 ≤ 2500 m∕s and 𝛥𝑡 ≤ 1.2 years.
Index No. of solutions MT-L1 MT-L2 TO-L1 TO-L2
MT-L1 /MT-L2 𝛥𝑉𝑚 (𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 )
/TO-L1 /TO-L2 m/s (day) day (m/s) m/s (day) day (m/s) m/s (day) day (m/s) m/s (day) day (m/s)
2 2878/5254/0/0 1189 (88) 11 (1558) 769 (136) 65 (1924) – – – –
3 2349/2829/0/0 941 (249) 97 (2463) 1284 (273) 46 (2443) – – – –
4 0/0/84/0 – – – – 2321 (346) 49 (2443) – –
5 1349/2172/0/0 2161 (344) 158 (2495) 2188 (337) 57 (2498) – – – –
6 0/0/61/54 – – – – 1575 (193) 116 (2069) 1679 (296) 43 (2208)
7 0/0/983/1077 – – – – 853 (204) 182 (971) 574 (184) 45 (782)
8 692/1810/0/0 2263 (326) 180 (2482) 2078 (313) 181 (2463) – – – –
9 0/0/275/616 – – – – 1331 (275) 11 (2009) 965 (294) 207 (2175)
10 0/1/46/47 – – 2320 (98) 98 (2320) 1483 (114) 83 (2012) 2153 (136) 136 (2153)
11 1159/3/0/0 1904 (315) 10 (2365) 1692 (39) 33 (2347) – – – –
12 0/0/32/127 – – – – 2237 (209) 124 (2375) 1545 (219) 131 (2378)
13 0/0/0/189 – – – – – – 1817 (55) 55 (1817)
14 0/0/100/144 – – – – 640 (122) 89 (753) 1168 (333) 186 (1562)
16 0/0/8/79 – – – – 2351 (221) 175 (2470) 1658 (289) 58 (2388)
17 0/0/427/506 – – – – 1313 (134) 130 (1330) 737 (288) 93 (2190)
18 0/0/752/840 – – – – 1503 (260) 239 (1673) 1323 (245) 96 (2493)
19 0/0/151/243 – – – – 2275 (245) 220 (2348) 1949 (360) 125 (2357)
20 1952/2880/0/0 1433 (280) 162 (1591) 1157 (71) 22 (1446) – – – –
21 0/0/46/21 – – – – 2028 (150) 119 (2438) 2014 (151) 150 (2285)
22 0/0/135/415 – – – – 2429 (172) 166 (2434) 2150 (27) 19 (2321)
23 2835/1947/0/0 1496 (31) 15 (2180) 897 (97) 62 (2392) – – – –
24 2137/0/0/0 1567 (331) 82 (2289) – – – – – –
25 0/0/41/89 – – – – 2200 (181) 167 (2283) 1460 (208) 100 (2305)
26 519/971/0/0 1798 (122) 102 (1821) 1720 (253) 67 (2353) – – – –
28 0/724/5/0 – – 1215 (56) 44 (2120) 2327 (280) 275 (2474) – –
29 0/78/0/0 – – 1978 (41) 34 (2434) – – – –
30 0/0/2/0 – – – – 2424 (310) 305 (2439) – –

Table 7
Characteristics of the best 3D rendezvous trajectories to the selected NEAs with 𝛥𝑉 ≤ 2500 m∕s and 𝛥𝑡 ≤ 1.2 years.
Index No. of solutions MT-L1 MT-L2 TO-L1 TO-L2
MT-L1 /MT-L2 𝛥𝑉𝑚 (𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 ) 𝛥𝑉𝑚 ∕(𝛥𝑡) 𝛥𝑡𝑚 ∕(𝛥𝑉 )
/TO-L1 /TO-L2 m/s (day) day (m/s) m/s (day) day (m/s) m/s (day) day (m/s) m/s (day) day (m/s)
31 3788/3578/0/0 1610 (338) 158 (1881) 929 (91) 59 (2342) – – – –
32 0/0/220/164 – – – – 1833 (180) 148 (2116) 1633 (286) 176 (2489)
34 0/0/0/2 – – – – – – 2471 (217) 217 (2474)
36 779/473/0/0 1805 (192) 190 (2041) 1211 (158) 94 (2488) – – – –
37 2911/5761/0/0 1575 (210) 203 (2250) 856 (225) 139 (1754) – – – –
38 0/3325/0/0 – – 1898 (123) 63 (2490) – – – –
39 0/0/27/157 – – – – 2392 (332) 332 (2392) 2188 (358) 358 (2216)
40 533/3670/0/0 2329 (166) 148 (2467) 1608 (109) 71 (2366) – – – –
41 1574/2204/0/0 1427 (123) 123 (1427) 802 (59) 43 (2051) – – – –
42 591/889/0/0 833 (232) 165 (876) 796 (202) 58 (1636) – – – –
43 3/353/0/0 2315 (182) 182 (2315) 1924 (260) 152 (2428) – – – –
44 0/0/1/3 – – – – 2419 (29) 29 (2419) 2126 (41) 29 (2133)
48 45/769/0/0 2146 (170) 162 (2387) 1566 (267) 23 (2196) – – – –
50 200/234/0/0 1180 (129) 129 (1180) 1419 (293) 267 (1925) – – – –
53 1/107/0/0 2485 (225) 225 (2485) 2100 (160) 58 (2459) – – – –
54 0/0/1/416 – – – – 2472 (343) 343 (2472) 1568 (326) 323 (1616)
55 0/0/0/1 – – – – – – 2324 (23) 23 (2324)
56 2/22/5/0 1892 (338) 336 (2337) 1945 (164) 140 (2477) 1566 (70) 36 (2018) – –
59 0/3/0/0 – – 2449 (218) 218 (2456) – – – –
61 11/0/0/0 1762 (303) 290 (2439) – – – – – –
62 0/0/1/1 – – – – 2435 (34) 34 (2435) 2219 (32) 32 (2219)
64 0/0/0/1 – – – – – – 1983 (23) 23 (1983)
65 0/0/0/2 – – – – – – 2078 (27) 27 (2078)
66 1/63/1/2 2481 (307) 307 (2481) 1840 (348) 334 (2495) 2143 (33) 33 (2143) 1356 (53) 29 (1543)
67 0/0/0/1 – – – – – – 2297 (24) 24 (2297)
69 1/0/0/0 2352 (280) 280 (2352) – – – – – –
70 0/0/0/1 – – – – – – 2050 (21) 21 (2050)
71 0/0/0/2 – – – – – – 2335 (60) 60 (2335)

maneuvers can be replaced with low-thrust arcs without altering the easily meet with the requirements set by NASA on the search for targets
transfer times (see [33]), reducing the propellant consumption due accessible by crewed missions [24].
to the higher specific impulse. The direct rendezvous trajectories de- The 3D trajectories from Earth to two neighboring asteroids via
scribed in this contribution can be viewed also as the building blocks transit orbits have been benchmarked against direct optimal patched-
of more complex mission scenarios, such as sample return operations conics transfers from a NASA study of accessible targets [23]. The
and asteroid tours. Furthermore, the 𝛥𝑉 budgets and transfer times test demonstrated the potential of the new method to find solutions

342
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

Fig. 16. Example of spatial rendezvous trajectories: ecliptic projection (left) and 3D view (right).

Table 8 Acknowledgments
Performance features of the direct rendevouz trajectories to the same two targets
computed by Strange et al. [23] and in this work.
EF, RF and GD acknowledge Khalifa University of Science and Tech-
Object 2003 YN107 (index 4) 2006 QQ56 (index 6)
nology’s internal grant CIRA-2021-65/8474000413. EF and RF have
Source Ref. [23] This work Ref. [23] This work
been partially supported by Technology Innovation Institute through
𝛥𝑉 (m/s) 1250 2321 1885 1575
grant ELLIPSE/8434000533. EF acknowledges projects PID2020-112
𝛥𝑡 (day) 621 346 621 193
C3 (km2 /s2 ) 6.30 0.13 5.90 0.13 576GB-C21 and PID2021-123968NB-I00 of the Spanish Ministry of
Launch date 2019-Dec-19 2022-Mar-26 2020-Sep-02 2023-Mar-31 Science and Innovation.
Arrival date 2021-Aug-26 2023-Mar-06 2022-May-07 2023-Oct-11

References

[1] NASA Jet Propulsion Laboratory, Center for Near-Earth Objects Study - NEO
that compare favorably in terms of time or flight and/or propellant Basics, https://cneos.jpl.nasa.gov/about/basics.html. (Last Accessed 25 August
2023).
consumption relative to traditional approaches, while requiring only [2] L.A. McFadden, R.P. Binzel, Near-earth objects, in: L.-A. McFadden, P.R. Weiss-
modest computational resources. man, T.V. Johnson (Eds.), Encyclopedia of the Solar System, second ed., Elsevier,
The presentation only addressed targets close to Earth’s orbit. Nev- 2007, pp. 283–300.
ertheless, the accessible region can be expanded introducing bridging [3] N. Anthony, M.R. Emami, Asteroid engineering: The state-of-the-art of near-
earth asteroids science and technology, Prog. Aerosp. Sci. 100 (2018) 1–17,
interplanetary arcs, combining the exploration of near-Earth space with
https://doi.org/10.1016/j.paerosci.2018.05.001.
a reduction in the launch cost towards more distant objects. [4] J.M. Hedo, E. Fantino, M. Ruíz, J. Peláez, Minimum orbital intersection distance:
an asymptotic approach, Astron. Astrophys. 633 (2020) A22, https://doi.org/10.
1051/0004-6361/201936502.
CRediT authorship contribution statement
[5] J.S. Lewis, Mining the Sky: Untold Riches from the Asteroids, Comets, and
Planets, Addison-Wesley Publishing Company, Reading (MA), 1997.
E. Fantino: Writing – review & editing, Writing – original draft, [6] L. Prockter, S. Murchie, A. Cheng, et al., The NEAR shoemaker mission to
Software, Resources, Project administration, Methodology, Investiga- asteroid 433 eros, Acta Astronaut. 51 (1) (2002) 491–500, https://doi.org/10.
1016/S0094-5765(02)00098-X.
tion, Funding acquisition, Formal analysis, Conceptualization. R. Flo-
[7] W.H. Blume, Deep impact mission design, Space Sci. Rev. 117 (2005) 23–42,
res: Writing – review & editing, Methodology, Investigation, Conceptu- https://doi.org/10.1007/s11214-005-3386-4.
alization. G. Donnarumma: Writing – review & editing, Visualization, [8] M. Yoshikawa, J. Kawaguchi, A. Fujiwara, A. Tsuchiyama, The hayabusa mission,
Software, Methodology, Investigation. D. Canales: Writing – review in: A. Longobardo (Ed.), Sample Return Missions, Elsevier, 2021, pp. 123–146,
& editing, Methodology. K.C. Howell: Writing – review & editing, https://doi.org/10.1016/B978-0-12-818330-4.00006-9.
[9] M.-K.J. Chung, B. Shyamkumar, S.R. Chesley, et al., EPOXI trajectory and ma-
Methodology.
neuver analyses, in: AAS/AIAA Astrodynamics Specialist Conference, Girdwood
(AK), 2011, Paper AAS 11-482.
Declaration of competing interest [10] L. Liu, Y. Lie, J.-F. Cao, et al., Mission design of the chang’e-2 asteroid
exploration mission, J. Astronaut. 35 (2014) 262–268, https://doi.org/10.3873/
j.issn.1000-1328.2014.03.003.
The authors declare that they have no known competing finan- [11] Y. Tsuda, T. Saiki, F. Terui, et al., Hayabusa 2 mission status: landing, roving,
cial interests or personal relationships that could have appeared to and cratering on asteroid ryugu, Acta Astronaut. 171 (2020) 42–54, https:
influence the work reported in this paper. //doi.org/10.1016/j.actaastro.2020.02.035.

343
E. Fantino et al. Acta Astronautica 229 (2025) 333–344

[12] D.S. Lauretta, S.S. Balram-Knutson, E. Beshore, et al., OSIRIS-REx: Sample return [24] P.A. Abell, B.W. Barbee, R.G. Mink, et al., The near-earth object human space-
from asteroid (101955) bennu, Space Sci. Rev. 212 (2017) 925–984, https: flight accessible targets study (NHATS) list of near-earth asteroids: Identifying
//doi.org/10.1007/s11214-017-0405-1. potential targets for future exploration, in: 43𝑟𝑑 Lunar and Planetary Science
[13] B. Williams, P. Antreasian, E. Carranza, et al., OSIRIS-REx flight dynamics and Conference, The Woodlands (The Netherlands), 2012, Paper 2842.
navigation design, Space Sci. Rev. 214 (69) (2018) https://doi.org/10.1007/ [25] J.P. Sánchez, D.Y. García, Asteroid retrieval missions enabled by invariant
s11214-018-0501-x. manifold dynamics, Acta Astronaut. 127 (2016) 667–677, https://doi.org/10.
[14] A.S. Rivkin, A.F. Cheng, Planetary defense with the double asteroid redirection 1016/j.actaastro.2016.05.034.
test (DART) mission and prospects, Nature Commun. 14 (2023) 1003, https: [26] A. Jorba, B. Nicolás, Using invariant manifolds to capture an asteroid near the
//doi.org/10.1038/s41467-022-35561-2. L3 point of the earth-moon bicircular model, Commun. Nonlinear Sci. Numer.
[15] D. Dunham, J. McAdams, R. Farquhar, NEAR mission design, in: Johns Hopkins Simul. 102 (2021) 105948, https://doi.org/10.1016/j.cnsns.2021.105948.
APL Technical Digest (Applied Physics Laboratory), Vol. 1, 2002, pp. 18–33. [27] E. Fantino, R. Castelli, Efficient design of direct low-energy transfers in multi-
[16] A.F. Cheng, A.S. Rivkin, P. Michel, J. Atchison, O. Barnouin, L. Benner, N.L. moon systems, Celestial Mech. Dynam. Astronom. 127 (2017) 429–450, https:
Chabot, C. Ernst, E.G. Fahnestock, M. Kueppers, P. Pravec, E. Rainey, D.C. //doi.org/10.1007/s10569-016-9733-9.
Richardson, A.M. Stickle, C. Thomas, AIDA dart asteroid deflection test: Planetary [28] D. Canales, K.C. Howell, E. Fantino, Transfer design between neighborhoods of
defense and science objectives, Planet. Space Sci. 157 (2018) 104–115, https: planetary moons in the circular restricted three-body problem: the moon-to-moon
//doi.org/10.1016/j.pss.2018.02.015. analytical transfer method, Celestial Mech. Dynam. Astronom. 133 (36) (2021)
[17] D.D. Mazanek, R.G. Merrill, J.R. Brophy, R.P. Mueller, Asteroid redirect mission https://doi.org/10.1007/s10569-021-10031-x.
concept: A bold approach for utilizing space resources, Acta Astronaut. 117 [29] D. Canales, K.C. Howell, E. Fantino, A versatile moon-to-moon transfer design
(2015) 163–171, https://doi.org/10.1016/j.actaastro.2015.06.018. method for applications involving libration point orbits, Acta Astronaut. 198
[18] L. Casalino, G. Colasurdo, Missions to asteroids using solar eletric propusion, Acta (2022) 388–402, https://doi.org/10.1016/j.actaastro.2022.06.010.
Astronaut. 50 (1) (2002) 705–711, https://doi.org/10.1016/S0094-5765(02) [30] D. Canales, K.C. Howell, E. Fantino, A.J. Gilliam, Transfers between moons with
00006-1. escape and capture patterns via Lyapunov exponent maps, J. Guid. Control Dyn.
[19] D.P.S. Santos, A.F.B.A. Prado, L. Casalino, G. Colasurdo, Optimal trajectories 46 (11) (2023) 2133–2149, https://doi.org/10.2514/1.G007195.
towards near-earth-objects using solar electric propulsion (SEP) and gravity [31] G. Donnarumma, Efficient Design of Optimal Low-Energy Trajectories to Near-
assisted maneuver, in: 32◦ Congresso Nacional de Matemática Aplicada e Earth Objects (Master’s thesis), Università degli Studi di Napoli Federico II,
Computacional, Culabá (Brazil), 2009, pp. 860–866. Department of Industrial Engineering, 2022.
[20] Z. Hasnain, C.A. Lamb, S.D. Ross, Capturing near-earth asteroids around earth, [32] E. Fantino, G. Donnarumma, M. Grassi, Efficient design of optimal of low-
Acta Astronaut. 81 (2) (2012) 523–531, https://doi.org/10.1016/j.actaastro. energy trajectories to near-earth objects, in: 2𝑛𝑑 International Stardust Conference
2012.07.029. (STARCON 2), ESA ESTEC, November 7 - 11, 2022.
[21] D. Landau, J. Dankanich, N. Strange, et al., Trajectories to nab a NEA (near-earth [33] D. Canales, E. Fantino, K.C. Howell, R. Flores, Design of optimal spatial low-
asteroid), in: AAS/AIAA Spaceflight Mechanics Meeting, Kauai (HI), 2013, Paper energy trajectories to near-earth objects, in: 74𝑡ℎ International Astronautical
AAS 13-409. Congress, Baku (Azerbaijan), 2023, Paper IAC-23-C1.6.1.
[22] Mascolo, A.D. Iuliis, L. Casalino, Fast and accurate estimation of fuel-optimal [34] NASA Jet Propulsion Laboratory, Astrodynamic Parameters, https://ssd.jpl.nasa.
trajectories to near-earth asteroids , Acta Astronaut. 188 (2018) 49–56, https: gov/astro_par.html. (Last Accessed 25 August 2023).
//doi.org/10.1007/s42064-018-0024-y. [35] T.S. Parker, L. Chua, Practical Numerical Algorithms for Chaotic Systems,
[23] N. Strange, E. Asphaug, K. Grogan, et al., Planetary Science Decadal Survey Springer-Verlag, New York, 1989.
- Near-Earth Asteroid Trajectory Opportunities in 2020–2024, Tech. Rep., Jet [36] NASA Jet Propulsion Laboratory, Small Bodies Database Query, https://ssd.jpl.
Propulsion Laboratory, California Institute of Technology, 2010. nasa.gov/tools/sbdb_query.html. (Last Accessed 25 August 2023).
[37] W.L.S. Wen, A study of cotangential, elliptical transfer orbits in space flight, J.
Aerosp. Sci. 28 (5) (1961) 411–417, https://doi.org/10.2514/8.9010.

344