Initial Guess Generation for Low-Thrust Trajectory Design with Robustness to Missed-Thrust-Events
Amlan Sinha, Ryne Beeson · 2025
Read Initial Guess Generation for Low-Thrust Trajectory Design with Robustness to Missed-Thrust-Events by Amlan Sinha, Ryne Beeson (2025) 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-2501-06694.txt.
Introduction
Amlan Sinha and Ryne Beeson present 2025 work on first guesses for low-thrust space routes that must survive missed thrust events. Interest in the space near the Moon has lifted demand for lean missions that thrust gently for long arcs. Those long arcs are prone to missed burns from faults or pauses, so early design must bake in robustness while plans still shift fast.
The hard part is that robust design solves large nonlinear problems, which need good first guesses to converge at all. The team compares a plain global search, which samples from one fixed spread, with a conditional search. The second builds guesses from solutions to less robust versions of the same problem, step by step. In a case study for a lunar gateway element, the conditional route converged far more often and to better routes.
Catalog record
| CREATOR | Amlan Sinha, Ryne Beeson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 97 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-06694
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Initial Guess Generation for Low-Thrust Trajectory Design with Robustness to Missed-Thrust-Events Amlan Sinha, Ryne Beeson Abstract The growing interest in cislunar space exploration in recent years has driven an increasing demand for efficient low-thrust missions to key cislunar orbits. These missions, typically possessing long thrust arcs, are particularly susceptible to operational uncertainties such as missed thrust events. Addressing these challenges requires efficient robust trajectory design frameworks during the preliminary mission design phase, where it is necessary to explore the solution space at a rapid cadence under evolving operational constraints. However, existing methods for missed thrust design rely on solving high-dimensional nonlinear programs, where generating effective initial guesses becomes challenging. To enhance computational efficiency, quality, and depth of robustness of solutions from global search, we compare two initial guess strategies: a baseline non-conditional global search, which samples from a static distribution with global support, and a conditional global search, which generates initial guesses conditioned on solutions to problems with less depth of robustness. The conditional search provides a sequential procedure for solving increasingly robust problems. We validate the improvements in the conditional approach using a low-thrust case study for the Lunar Gateway Power and Propulsion Element, where our results demonstrate that it significantly improves convergence rate and solution quality, highlighting its potential in preliminary robust trajectory design. Nomenclature \Omega = random sample space \mathcal{F} = \sigma -algebra of measurable events (\mathcal{F}_{t})_{t\geq 0} = filtration representing information over time \mathbb{P} = probability measure \omega = random variable J = objective function \phi = terminal cost \mathcal{L} = running cost \bm{f} = uncontrolled state vector field \bm{g} = controlled state vector field t_{0} = initial time t_{f} = final time \bm{\xi} = spacecraft state \bm{u} = spacecraft control \bm{x} = nonlinear programming decision variable \bm{c} = nonlinear programming constraint T_{s} = shooting time T_{i} = initial coast time T_{f} = final coast time \mathcal{E} = equality index set \mathcal{I} = inequality index set N = number of segments \overline{N} = number of decision variables \tau_{1} = time where a missed thrust event begins \Delta\tau = duration of the missed thrust event \mathcal{X} = solution space X = initial guess generator \mathcal{M} = mapping strategy \mathcal{S}(k) = Non-conditional strategy for a robust problem with k realizations \mathcal{S}(k|k^{\prime}) = Conditional strategy for a k -robust problem with initial guesses derived from a k^{\prime} -robust problem. Superscripts \dagger = reference solution \omega = realization solution Subscripts i,j = indices k,k^{\prime} = number/index of realizations 1 Introduction I n the recent past, low-thrust (LT) propulsion systems are becoming increasingly popular for space missions, particularly cislunar. Several Artemis I payloads, such as Lunar IceCube, EQUULEUS, Near-Earth Asteroid Scout, BioSentinel, LunIR, and Lunar Polar Hydrogen Mapper, exemplify this trend. Despite their many advantages, LT spacecrafts are particularly susceptible to safe mode events , which can occur if an anomalous event (e.g., impact with space debris) forces the spacecraft to enter a protective mode during which all thruster operations are switched off. If such an event occur coincides with a scheduled thrust arc, it results in what is known as a missed thrust event (MTE). Due to their characteristically long thrust arcs, MTEs are relatively common in LT missions [ 1 ] . If not adequately addressed during the preliminary design phase, they can severely compromise mission performance and, in certain cases, lead to complete failure - especially if they disrupt maneuvers which must occur at critical junctures along the trajectory (e.g., flybys). (a) Time Since 2026/04/26 14:47:58 [days] Wet Mass [kg] Throttle (b) Figure 1 : Reoptimizing a nominal solution following an MTE can deteriorate the baseline As a motivating example, consider a LT spacecraft, modeled after the Power and Propulsion Element [ 2 , 3 ] (discussed in more detail in § 5 ), on a spiral trajectory in the cislunar realm, where the objective is to identify a feasible, minimum fuel transfer to the Earth–Moon \mathcal{L}_{2} Southern 9:2 Near Rectilinear Halo Orbit. Due to the low thrust acceleration, solutions to this problem typically span multiple revolutions with long thrust arcs, making them particularly vulnerable to MTEs. For illustration, let us consider a nominal solution which consumes approximately 2,038 kg of propellant over a total flight time of 225 days (Fig. 1 ). To understand how an MTE could affect this solution, we simulate a 60-hour outage near the beginning of the transfer, preventing the spacecraft from executing the first scheduled burn, and forcing it to coast under natural dynamics until it can resume its thruster operations once again. Reoptimizing this solution after the recovery yields a trajectory which differs significantly from the original. Figure 1(a) illustrates the nominal non-robust trajectory in green alongside the reoptimized trajectory in pink . The LT spiral phase is highlighted in beige , the Moon in grey , and the NRHO in dark blue . During the MTE, the spacecraft is forced to coast ballistically, shown with a dashed light blue line. Figure 1(b) displays the corresponding wet mass and throttle profiles for the post-spiral phase in both solutions (nominal profiles in green ; reoptimized profiles in pink ), with dashed lines indicating average throttle levels in each case. The reoptimized trajectory deviates significantly from the nominal, initially trailing behind due to the MTE. It quickly closes the gap, ultimately completing the mission 5 days earlier than the nominal, but this, however, increases the propellant requirement by 8 kilograms. The corresponding throttle profiles reveal that while both solutions follow similar patterns, the reoptimized solution exhibits a slightly higher average throttle, by approximately 10%, in comparison to the nominal solution. Although a viable solution was found in this case, such recovery may not always be possible. The spacecraft’s ability to recover from such engine failures depends on several factors, including its current state, the available on-board propellant, and the remaining mission time, underscoring the importance of designing reliable contingencies during the preliminary mission design phase. Doing so will allow us to not only build in the necessary margins but also ensure that an alternative solution is readily available to deploy in the event that such outages disrupt the execution of a scheduled burn. In the most general case, a robust optimal control problem involves designing control strategies to ensure feasibility or near-optimality in the presence of uncertainties, which can be broadly classified as aleatoric or epistemic. While the robust optimal control problem is of interest to the broader controls community, it has gained particular traction in the aerospace domain, where aleatoric uncertainties (e.g., navigational errors), epistemic uncertainties (e.g., imperfect knowledge about system/spacecraft parameters), and maneuver execution errors can significantly impact mission success. In practice, robustness is typically evaluated a posteriori through navigation analysis during the preliminary mission design phase, followed by iterative adjustments to the nominal solution. Common methods to improve robustness include the addition of empirical margins, reductions in duty cycles, or incorporation of coasting arcs along sensitive solution segments. However, these adjustments, which are often decoupled from the optimization process, can lead to suboptimal solutions with overly conservative margins, reducing overall mission efficiency. Existing approaches to robust optimal control generally address these uncertainties by either reformulating the problem to account for worst-case adversarial uncertainty realizations or by considering a range of