Reachability-Based Safe-Start Regions for Approach to a Tumbling Target with Rotating LOS Constraints

Erick Lansard; Keck Voon Ling; Omer Burak Iskender; Wee Seng Lim

REVIEW 4 major objections 6 minor 21 references

Reachability-Based Safe-Start Regions for Approach to a Tumbling Target with Rotating LOS Constraints

T0 review · 4 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Two closed-form inequalities certify safe starting states for approach to a tumbling target, giving a set strictly contained in the numerical reachable set and computed about 250 times faster.

desk verdict A genuinely useful empirical benchmark and a sensible conceptual distinction, but the 'sound inner certificate' claim is not established: the closed-form criteria are unproved kinematic bounds, the paper's own Table 3 shows 30 false positives, and Section 7.2 contradicts the abstract. read the letter →

arxiv 2607.02128 v3 pith:HSQYTZK2 submitted 2026-07-02 eess.SY cs.SYphysics.space-ph

Omer Burak Iskender , Keck Voon Ling , Wee Seng Lim , Erick Lansard This is my paper · ORCID

classification eess.SYcs.SYphysics.space-ph MSC 93B0370M20

keywords spacecraftrendezvoustumblingtargetline-of-sightcorridorbackwardreachablesetsynchronizationradiusHamilton-Jacobireachabilitymodelpredictivecontrolgo/no-gocertification

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks: from which initial states can a spacecraft safely approach a tumbling, uncooperative target while staying inside a line-of-sight docking corridor that rotates with the target's body frame? The authors derive a closed-form answer from bounded-thrust relative orbital dynamics: each corridor constraint is eroded by a margin δ_i = (sdot_i^-)^2/(2 a_max) to account for rotation-induced drift before thrust can arrest it, and the chaser must start within a synchronization radius r < 2 a_max/ω_t^2 over which the apparent rotational velocity can be cancelled. They argue this synchronization set is a sound inner certificate — strictly smaller than the positional reachable set computed by Hamilton-Jacobi reachability, because requiring co-rotation is stronger than merely arriving at the hold point. Across a 500-case sweep with a closed-loop MPC controller, the certificate predicts empirical feasibility with 0.80 precision and 0.91 recall at a fraction of the compute cost, so the authors propose it as an onboard go/no-go bound for pre-mission planning and replanning.

What carries the argument

Two analytical objects do the work. First, the directional per-constraint erosion margin δ_i = (sdot_i^-)^2 / (2 a_max): for each face of the rotating LOS corridor, this is the distance the constraint slack erodes under apparent rotational drift before the thruster at acceleration a_max can arrest it; the initial state is safe only if slack exceeds this margin on every face. Second, the synchronization radius r_sync = 2 a_max / ω_t^2, obtained by requiring that the braking distance ω_t^2 r^2/(2 a_max) to cancel the apparent co-rotation velocity ω_t r be smaller than the range r itself. Together they form a closed-form certificate that the chaser can brake into co-rotation with the hold point

What would settle it

Run the exact closed-loop dynamics (Clohessy-Wiltshire-Hill with sub-stepping, no blending or projection) from a state that passes conditions (8)-(9) near the boundary of the synchronization radius, such as at a_max=0.10 m/s^2, tumble rate 4 deg/s, range 30 m where the paper reports 7/9 certified but 0/9 feasible; if the corridor is violated or co-rotation is not achieved in that test, the claimed soundness is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the safe-start question for approach to a tumbling target under a rotating line-of-sight corridor splits into two distinct notions: positional reachability (whether the chaser can reach the hold point at all, with any terminal velocity) and synchronization (whether it can both reach the hold point and match its co-rotation velocity ω_t r). The authors derive a closed-form synchronization set from two inequalities — the per-constraint directional erosion δ_i = (sdot_i^-)^2/(2a_max) and the synchronization radius r < 2a_max/ω_t^2 — and show, by comparison with Hamilton-Jacobi backward reachable sets, that this set is strictly contained in the positional reacha

Load-bearing premise

The certificate assumes the chaser's full thrust authority can be applied independently along each corridor constraint and that simply braking the apparent rotational velocity over the range r is enough to co-rotate with the hold point; it never proves this remains feasible under the CWH orbital coupling (the 3n^2x, 2n ydot, -2n xdot terms) and the sustained centripetal acceleration ω_t^2 r needed for co-rotation.

Editorial extensions

If this is right

If the certificate is correct, a spacecraft can run a ~10 ms check on an initial state and decide go/no-go for approach to a tumbling target without solving a 4D Hamilton-Jacobi PDE (which takes ~2.4 s per query).
The strict containment S_sync ⊊ S_reach implies that any mission requiring sustained co-rotation (docking, berthing, capture) must plan from a strictly smaller set than mere reachability would suggest; the gap quantifies the extra cost of insisting on synchronization.
The ∝1/ω_t^2 collapse of r_sync means high tumble rates make co-rotation infeasible beyond very short ranges even with high thrust authority; this gives a mission-design rule for choosing approach geometry or deciding to despin the target.
Because the certificate is conservative by construction (a sound inner bound), a state that passes is provably safe for the modeled dynamics; states that fail may still be feasible, as the 12 false negatives in the sweep show.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The erosion formula treats each corridor constraint axis independently and ignores the CWH coupling terms and the centripetal acceleration needed to stay co-rotating. A natural extension would be to tighten δ_i to account for the coupling, or to augment the certificate with a tube based on the CWH dynamics; the paper's own 6% false positives at high tumble rates (concentrated at the handover) hint
The synchronization radius r_sync = 2a_max/ω_t^2 has the same form as a classic acceleration-limited stopping distance, but expressed in angular terms; one could test the conjecture that the true synchronization set in CWH coordinates is obtained by replacing a_max with the minimum singular value of the input map along the co-rotation trajectory, which would give a dynamic, state-dependent version
The confusion-matrix numbers (122 TP, 30 FP, 12 FN) suggest that as a go/no-go screen the certificate is high-recall but not sufficient; a hierarchical scheme could use the 10 ms test for screening and a short-horizon reachability or control-barrier check only for states in the 6% ambiguous band.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here's the short version: this paper has a useful empirical study and a correct qualitative point, but the headline claim of a "sound inner certificate" doesn't hold up. The closed-form criteria (8)-(9) are two braking-distance inequalities from a double integrator, applied to a CWH problem with a rotating corridor, and no proof connects them to feasibility of the actual dynamics. The paper's own Section 7.2 says "not a sufficient certificate," while the abstract and conclusions call it a sound inner bound. That's a direct contradiction on the central safety claim.

What's genuinely good: the benchmarking against four reachability engines is careful, and the decision to use CWH truth propagation without reference blending or state projection is the right call — it exposes real physical limits instead of phantom successes. The confusion matrix in Table 3 is transparent, and the structural observation that synchronization (reach the hold point and co-rotate) is strictly stronger than positional reachability is correct and worth saying. The gap widening with tumble rate is intuitive and the sweeps confirm it.

The soft spots are load-bearing. The erosion margin (8) assumes each corridor constraint can be arrested with full thrust independently. The synchronization radius (9) assumes co-rotation is achieved merely by braking the apparent velocity over the available range. Neither accounts for the CWH coupling terms in (1) nor for the sustained centripetal acceleration needed to track a rotating point. A concrete check: at ω_t = 4 deg/s and r = 30 m, the centripetal acceleration is ω_t² r ≈ 0.146 m/s², which exceeds the a_max = 0.10 m/s² used in Table 3. Yet r_sync = 2 a_max / ω_t² ≈ 41 m, so criterion (9) certifies that cell. That's not a controller artifact; it's a physical impossibility for the exact synchronization task. The inclusion chain (12) is asserted, not derived. The 30 false positives in Table 3 may be partly due to the single-input filter's projection and inter-sample violations, but the paper hasn't shown they are entirely due to those, and the centripetal argument suggests they are not.

Who this is for: people working on proximity operations with tumbling targets, especially those who want a quick screening test and a comparison of reachability tools. It deserves a serious referee because the problem matters and the empirical data is honestly reported. But I would not accept the safety claim as stated. A revision needs to either prove sufficiency of (8)-(9) for the CWH + rotating-corridor dynamics (including the centripetal constraint) or explicitly downgrade the claim to a screening heuristic and change the abstract and conclusions accordingly. The omitted high-tumble cell in Table 3 should also be filled in. Send it to peer review, but expect major revision.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes two closed-form safe-start criteria for approach to a tumbling target under a rotating LOS corridor: a directional per-constraint erosion margin, δ_i = (sdot_i^-)^2/(2 a_max) (Eq. 8), and a synchronization radius, r < 2 a_max / ω_t^2 (Eq. 9). These are embedded in a three-regime PD/MPC guidance architecture and benchmarked against backward/forward polytopic reachable sets, Hamilton–Jacobi reachability, and closed-loop Monte Carlo. The central claim is that the closed-form synchronization set is a sound inner certificate, strictly contained in the positional HJ reachable set (S_sync ⊊ S_reach), and that it can serve as a ~10 ms onboard go/no-go test where HJ requires seconds.

Significance. A validated ~10 ms sufficient certificate for safe starts would be a practically valuable contribution to on-orbit servicing and debris removal, and the paper has genuine strengths: it ships reproducible open-source code, uses physically honest CWH truth propagation without state projection or reference blending, and benchmarks against four independent reachability engines over a parametric sweep. However, the advertised certificate claim is not supported. The analytical derivation is double-integrator braking geometry, not a consequence of the CWH dynamics stated in Eq. (1), and the paper's own Table 3 shows 30 certified-but-infeasible cases (precision 0.80), which is incompatible with a sound inner certificate. As an empirical high-recall screening heuristic the results may be useful, but the manuscript as written overclaims its central contribution, and the internal contradiction between the Conclusions and §7.2 is load-bearing.

major comments (4)

[§5.1–5.2, Eqs. (8)–(9)] The closed-form criteria are derived from scalar double-integrator braking geometry, not from the CWH dynamics (1). Eq. (8) credits the full thrust a_max independently to each constraint slack, and Eq. (9) uses the braking distance d_brake = ω_t^2 r^2/(2 a_max) without accounting for the sustained centripetal acceleration ω_t^2 r needed to co-rotate with the hold point or the CWH coupling terms (3 n^2 x, 2 n ydot, -2 n xdot). No theorem states that a state satisfying (8)–(9) is feasible for the dynamics (1) under the rotating corridor. Since the abstract claims these are 'derived from bounded-thrust relative orbital dynamics,' this is a load-bearing gap. The authors should either prove sufficiency under explicit assumptions or abandon the 'sound inner certificate' wording.
[§7.2, Table 3] The numerical sweep directly contradicts the 'sound inner certificate' claim. The confusion matrix reports TP=122, FP=30, TN=336, and the text notes precision 0.80. For example, at a_max=0.10 m/s^2, ω_t=4 deg/s, r0=30 m, the table gives 7/9 certified but 0/9 feasible. A sufficient inner certificate must have zero false positives; 30 certified-but-infeasible cases are concrete counterexamples. The paper itself concedes in §7.2 that the criteria are 'not a sufficient certificate.' This is not a presentation issue; it invalidates the abstract/conclusions claim that the criteria are a sound inner bound.
[§7.1, Eq. (12)] The inclusion chain S_tube ⊆ S_ana ⊆ S_sync ⊆ S_pos_HJ is asserted, not derived. In particular, S_ana ⊆ S_sync requires that every analytically certified state can both reach and co-rotate with the hold point; no such proof is given, and Table 3 suggests the analytical set over-approximates the empirically synchronization-feasible set. The HJ benchmark is a terminal-time position reach set, not a reach-and-co-rotate set, so the IoU numbers in Table 2 do not validate Eq. (12). Each inclusion should be proved under stated assumptions or else explicitly labeled as conjecture.
[Conclusions vs §7.2] There is an internal contradiction that cannot be ignored: the Conclusions call (8)–(9) a 'sound inner certificate,' while §7.2 states they are 'not a sufficient certificate' and the confusion matrix shows false positives. Because the operational value of the paper is the claimed onboard go/no-go guarantee, this is not a local wording issue. The manuscript should be reframed around the actual result: a high-recall, low-cost screening heuristic with a quantified false-positive rate, not a certified safe-start set.

minor comments (6)

[§5.1, Eq. (8)] The constraint slack s_i and its derivative sdot_i are used before being defined. Please define s_i = b_c,i - A_c,i R_z(-θ) r and give the explicit expression for sdot_i under the rotation.
[§4.2] The QP enforces the axis-aligned box |u_i| ≤ a_max and then projects to the Euclidean disk ∥u∥_2 ≤ a_max. This post-solve projection can destroy optimality and does not guarantee the QP constraints remain feasible. Clarify whether the projection is part of the controller or a heuristic that contributes to the false positives discussed in §7.2.
[Table 3] The table caption says 'Each cell totals 9 unless otherwise noted,' but the 0/8 cell at a_max=0.02, ω_t=1, r0=150 is noted, and the text says it provides 8 of the 12 false negatives. The total of 500 cases requires a full accounting of missing cases; please give an explicit count of cases per cell.
[§3.2, Eq. (5)] The body-frame rotation is about the LVLH z-axis at rate ω_t, but the tumble of an arbitrary target is generally three-axis. Please clarify whether the analysis is restricted to spin about the orbit normal and state this limitation in the Introduction or problem setup.
[References] References [16] and [18] appear to be concurrent submissions to the same proceedings (IAC 2026). If they are not yet published, please provide status or preprint DOIs so the reader can verify the relevant claims; if they are self-citations to work under review, this should be disclosed.
[§8.3] The sentence 'The precision 0.80/recall 0.91 closed-loop agreement ... supports the use of r_sync and the directional-erosion test as onboard go/no-go signals' is too strong given the 30 false positives. A go/no-go signal with 6% false positives may be acceptable for screening, but it should be described as screening, not as a safety certificate.

Circularity Check

0 steps flagged · score 0.0 of 10

No material circularity: the analytical safe-start criteria are independent kinematic inequalities benchmarked against independent HJ and Monte Carlo engines; self-citations are contextual and not load-bearing.

full rationale

The closed-form criteria are derived directly from the stated kinematic quantities: the per-constraint erosion margin (8) is the scalar braking distance (sdot_i^-)^2/(2 a_max), and the synchronization radius (9) is the algebraic consequence of d_brake = omega_t^2 r^2/(2 a_max) < r. Neither is fitted to, or read off, the Hamilton-Jacobi, polytopic, or Monte Carlo results. The benchmark in Section 7 compares these analytical sets against independent engines and reports IoU values and a confusion matrix that could have falsified the criteria; indeed Table 3 reports 30 false positives, which the paper does not hide. The strict-containment claim S_sync subset S_reach is a definitional relation between synchronization (reach and co-rotate) and positional reachability, not a fitted prediction, and the paper explicitly defines the HJ target as position-only. Self-citations such as [10], [14]-[16], and [18] are contextual, tutorial, or confined to planned extensions; none supplies the load-bearing soundness of (8)-(9). The genuine weakness, that the soundness of (8)-(9) with respect to the full CWH dynamics is asserted via the inclusion chain (12) rather than proved, and that Section 7.2 concedes the criteria are 'not a sufficient certificate' while the Conclusions call them 'a sound inner certificate', is an internal-support or correctness issue, not a circularity of the kind this pass is asked to flag.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The analytical criteria themselves have no fitted constants, but they rest on two unproved ad-hoc kinematic assumptions (independent full-authority braking and no centripetal maintenance) plus the CWH and HJ modeling assumptions. MPC weights and horizons are hand-chosen and directly affect the empirical false-positive count.

free parameters (3)

MPC weights w_u and w_delta_u = w_u = 0.1, w_delta_u = 2.0
Chosen by hand; affect the Table 3 closed-loop feasibility labels but not the analytical criteria.
MPC horizon N and sampling interval dt = N = 6, dt = 2 s
Fixed for the safety filter; the paper itself identifies horizon discretization as a cause of inter-sample LOS violations.
Far/close handover ramp budget T_ramp = 10 dt
Tuning gain for the three-regime controller; deliberately distinct from the analytical synchronization radius.

assumptions (5)

domain assumption CWH linearized relative dynamics are an adequate truth model at the ranges and tumble rates considered (up to ~150 m).
All feasibility claims are evaluated with CWH truth propagation (Section 2.1), so errors in CWH linearization transfer to the empirical labels.
domain assumption The rotating LOS corridor is exactly a fixed polyhedral cone in the target body frame rotating about LVLH z at constant rate omega_t.
Used in (4) and throughout Sections 3-7; real tumble is three-dimensional and not constant-rate.
ad hoc to paper Full thrust authority a_max can be applied independently in the direction of each constraint slack, with no thrust-sharing or centripetal coupling.
This is the unproved sufficient-condition assumption behind (8) and the 'sound inner certificate' claim.
ad hoc to paper Co-rotation is achieved by stopping the apparent rotational velocity over the available range; no sustained centripetal acceleration is required.
Underlies (9); ignores the omega_t^2 r acceleration needed to maintain a synchronized hold.
domain assumption HJ level-set solution on a 31x31x15x15 grid with a terminal-time target is an adequate ground truth for positional reachability.
Used as the reference for IoU comparisons in Table 2; grid dissipation and terminal-only target affect the comparison.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Reachability-Based Safe-Start Regions for Approach to a Tumbling Target with Rotating LOS Constraints." pith.science (2026). https://pith.science/paper/HSQYTZK2

@misc{pith2026260702128,
  author       = {Pith},
  title        = {Pith review of: Reachability-Based Safe-Start Regions for Approach to a Tumbling Target with Rotating LOS Constraints},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HSQYTZK2}},
  note         = {Machine review of arXiv:2607.02128}
}

read the original abstract

This paper presents a reachability-aware guidance architecture for autonomous approach to a tumbling, uncooperative target under a rotating line-of-sight (LOS) docking corridor. The LOS admissible set rotates with the target body frame, producing time-varying polyhedral constraints on the chaser's relative motion, and the design problem is to establish the initial states from which the chaser can safely approach and synchronize with the rotating hold point. A closed-form answer is derived from bounded-thrust relative orbital dynamics, in the form of two analytical safe-start criteria: a directional per-constraint erosion margin, quantifying the corridor margin consumed by rotation-induced drift before the thruster can arrest it, and a synchronization radius, bounding the range over which the apparent rotational velocity can still be cancelled. Closed-loop guidance combines a three-regime tracking law, spanning far-field approach, close body-frame tracking, and synchronized hold, with a receding-horizon quadratic program carrying the rotating corridor constraints and exact discrete relative-motion prediction. The criteria are benchmarked against four standard reachability engines, namely backward and forward polytopic reachable sets, Hamilton--Jacobi level sets, and closed-loop Monte Carlo simulation. The closed-form test is orders of magnitude cheaper than grid-based Hamilton--Jacobi reachability while tracking closed-loop feasibility closely across a parametric sweep. The residual optimism and the gap against Hamilton--Jacobi are structural rather than a method error: requiring the chaser to reach the hold point and co-rotate with it is strictly stronger than requiring it to arrive with arbitrary velocity, and the gap widens with tumble rate. The criteria therefore serve as an onboard go/no-go bound where Hamilton--Jacobi reachability is prohibitively expensive.

Figures

Figures reproduced from arXiv: 2607.02128 by the authors.

**Figure 2.** Planar CWH (𝑛 = 1.1 × 10−3 rad/s, blue solid hull) vs. a same-order double integrator (𝑛 = 0, grey dashed hull) backward reachable set for the three BRS variants; target hold [0, 30, 0, 0] ⊤, 𝑁 = 20, 𝑑𝑡 = 2 s, 𝑎max = 0.10 m/s2 . Orbital dynamics contract the radial (𝑥𝐵) reach to ∼81% of the double integrator’s (see text). 1. Far approach (𝑟 > 𝑟sync): LVLH-frame PD tracking of the spiral reference with velocity limi… view at source ↗

**Figure 3.** Combined safe-start region overview for six tumble rates ( [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 3.** Figure 3: Combined safe-start region overview for five tumble rates ( [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 1 linked inside Pith

[1]

Flores-Abad, O

A. Flores-Abad, O. Ma, K. Pham, and S. Ulrich. A review of space robotics technologies for on-orbit servicing.Progress in Aerospace Sciences, 68:1–26, 2014

2014
[2]

Researchesinthelunartheory.American Journal of Mathematics, 1(1):5–26, 1878

G.W.Hill. Researchesinthelunartheory.American Journal of Mathematics, 1(1):5–26, 1878
[3]

W. H. Clohessy and R. S. Wiltshire. Terminal guid- ance system for satellite rendezvous.Journal of the Aerospace Sciences, 27(9):653–658, 1960

1960
[4]

Borrelli, A

F. Borrelli, A. Bemporad, and M. Morari.Predictive Control for Linear and Hybrid Systems. Cambridge University Press, 2017

2017
[5]

Introduction to Model Predictive Control

F.Borrelli,M.Morari,andC.N.Jones. Introduction to Model Predictive Control. Lecture notes, ETH Zürich / UC Berkeley, 2015

2015
[6]

I.M.Mitchell,A.M.Bayen,andC.J.Tomlin.Atime- dependentHamilton-Jacobiformulationofreachable sets for continuous dynamic games.IEEE Transac- tions on Automatic Control, 50(7):947–957, 2005

2005
[7]

Bansal, M

S. Bansal, M. Chen, S. Herbert, and C. J. Tomlin. Hamilton-Jacobi reachability: A brief overview and recent advances. InProceedings of the 56th IEEE Conference on Decision and Control (CDC), pages 2242–2253, 2017

2017
[8]

Chen and C

M. Chen and C. J. Tomlin. Hamilton–jacobi reach- ability: Some recent theoretical advances and appli- cations in unmanned airspace management.Annual Review of Control, Robotics, and Autonomous Sys- tems, 1:333–358, 2018

2018

Show all 21 references

[9]

Girard and C

A. Girard and C. Le Guernic. Efficient reachability analysis for linear systems using support functions. IFACProceedingsVolumes,41(2):8966–8971,2008

2008
[10]

O. B. Iskender. Intelligent game countermeasures for non-cooperative spacecraft proximity operations on circular and elliptical orbits.Available at SSRN 6952642, 2026

2026
[11]

Sreenath, and P

A.D.Ames,S.Coogan,M.Egerstedt,G.Notomista, K. Sreenath, and P. Tabuada. Control barrier func- tions: Theory and applications. InProceedings of the18thEuropeanControlConference(ECC),pages 3420–3431, 2019

2019
[12]

AIAA Education Series, 4th edition, 2018

H.SchaubandJ.L.Junkins.AnalyticalMechanicsof Space Systems. AIAA Education Series, 4th edition, 2018

2018
[13]

D. Q. Mayne, M. M. Seron, and S. V. Raković. Robust model predictive control of constrained lin- ear systems with bounded disturbances.Automatica, 41(2):219–224, 2005

2005
[14]

O. B. Iskender, K. V. Ling, and V. Dubanchet. Con- straintstighteningapproachtowardsmodelpredictive control based rendezvous and docking with uncoop- erativetargets. InProceedingsofthe2018European Control Conference (ECC), pages 380–385, 2018

2018
[15]

O. B. Iskender.Model Predictive Control for Space- craft Rendezvous and Docking with Uncooperative Targets. PhD thesis, Nanyang Technological Univer- sity, Singapore, 2020

2020
[16]

O. B. Iskender and K. V. Ling. Horizon-dependent tube MPC for elliptical-orbit rendezvous under mass uncertainty. InProc. 77th Int. Astronautical Congr. Int. Astronautical Federation, 2026

2026
[17]

Schmerling, S

E. Schmerling, S. Herbert, and M. Pavone. hj_reachability: JAX-based Hamilton–Jacobi reachability toolbox.https://github.com/ StanfordASL/hj_reachability, 2021

2021
[18]

Sloshing awaretubeMPCwithpartialtighteningforspacecraft proximityoperations

O.B.IskenderandJ.L.RedondoGutiérrez. Sloshing awaretubeMPCwithpartialtighteningforspacecraft proximityoperations. InProc.77thInt.Astronautical Congr. Int. Astronautical Federation, 2026

2026
[19]

Simulation-basedLQR-trees with input and state constraints

P.ReistandR.Tedrake. Simulation-basedLQR-trees with input and state constraints. InProceedings of the IEEE International Conference on Robotics and Automation (ICRA), pages 5504–5510, 2010

2010
[20]

Regionofattractionestimationforfree-floatingLQR controllers

L.Maywald,S.Vyas,M.Ben-Larbi,andM.Janković. Regionofattractionestimationforfree-floatingLQR controllers. InProceedingsoftheGermanAerospace Congress (GVW-II), 2021

2021
[21]

Shala, S

E. Shala, S. Vyas, M. Ben-Larbi, M. Kumar, and E. Stoll. Region of attraction estimation for free- floating systems under time-varying LQR control. arXiv:2405.06726, 2024. 10

2024 arXiv

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [1]

Flores-Abad, O

A. Flores-Abad, O. Ma, K. Pham, and S. Ulrich. A review of space robotics technologies for on-orbit servicing.Progress in Aerospace Sciences, 68:1–26, 2014

2014

[2] [2]

Researchesinthelunartheory.American Journal of Mathematics, 1(1):5–26, 1878

G.W.Hill. Researchesinthelunartheory.American Journal of Mathematics, 1(1):5–26, 1878

[3] [3]

W. H. Clohessy and R. S. Wiltshire. Terminal guid- ance system for satellite rendezvous.Journal of the Aerospace Sciences, 27(9):653–658, 1960

1960

[4] [4]

Borrelli, A

F. Borrelli, A. Bemporad, and M. Morari.Predictive Control for Linear and Hybrid Systems. Cambridge University Press, 2017

2017

[5] [5]

Introduction to Model Predictive Control

F.Borrelli,M.Morari,andC.N.Jones. Introduction to Model Predictive Control. Lecture notes, ETH Zürich / UC Berkeley, 2015

2015

[6] [6]

I.M.Mitchell,A.M.Bayen,andC.J.Tomlin.Atime- dependentHamilton-Jacobiformulationofreachable sets for continuous dynamic games.IEEE Transac- tions on Automatic Control, 50(7):947–957, 2005

2005

[7] [7]

Bansal, M

S. Bansal, M. Chen, S. Herbert, and C. J. Tomlin. Hamilton-Jacobi reachability: A brief overview and recent advances. InProceedings of the 56th IEEE Conference on Decision and Control (CDC), pages 2242–2253, 2017

2017

[8] [8]

Chen and C

M. Chen and C. J. Tomlin. Hamilton–jacobi reach- ability: Some recent theoretical advances and appli- cations in unmanned airspace management.Annual Review of Control, Robotics, and Autonomous Sys- tems, 1:333–358, 2018

2018