Pith. sign in

REVIEW 4 major objections 5 minor 36 references

The paper's central claim is that a single linear constraint on the very first control input—requiring that the drone's future reachable set is never fully swallowed by a human's reachable set—recursively guarantees collision-free 3D naviga

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

The paper contributes a linear MPC safety constraint that, for the first control input alone, prevents the drone's future reachable set from ever being fully inside a human's reachable set, avoiding inevitable collisions.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection Genuinely novel MPC safety idea with solid experiments, but the central derivation has two concrete gaps; send to review so the theory gets fixed. the 4 major comments →

arxiv 2510.17525 v3 pith:O34MIQ6Z submitted 2025-10-20 cs.RO

HumanHalo -- Safe and Efficient 3D Navigation Among Humans via Minimally Conservative MPC

classification cs.RO
keywords safe robot navigationmodel predictive controlreachability analysiscollision avoidancemicro aerial vehiclehuman motion predictionzonotopesreal-time control
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

HumanMPC tries to show that formally safe drone navigation among moving humans is compatible with real-time, efficient flight. Instead of demanding the drone's entire trajectory stay outside the human's reachable region—the standard, conservative approach—it constrains only the control input executed now, while checking its consequences across the whole horizon. The safety constraint is linear, so the online problem stays a quadratic program that solves in milliseconds on embedded hardware. If the guarantees hold, this would let aerial robots interact closely with people in tasks like filming, inspection, or delivery without freezing or taking over-conservative detours. The paper validates the claim in simulation with real human motion data and on a live quadrotor.

Core claim

The core discovery is the noncontainment reachability constraint: at every horizon step, the drone's reachable set must not be a subset of the human's reachable set. If the drone's reachable set is ever fully inside the human's, collision is inevitable no matter what later controls are applied; if there is even one point outside, a future control can still escape. The authors derive a supporting-hyperplane condition that turns this geometric requirement into one linear inequality on the initial control input u0 (Eq. 18), and argue that satisfying this inequality recursively—step by step, in a receding-horizon MPC—guarantees feasibility and therefore safety, under stated assumptions about ini

What carries the argument

The central object is the linear safety constraint of Eq. (18), derived from zonotope reachable sets: the drone's reachable set is an exact zonotope from its linearized dynamics and box control bounds, while the human's is a conservative over-approximation built from capsules/spheres following the 24-joint body and an inflated cylinder. The support function of the zonotope gives a separating hyperplane, and substituting the stacked dynamics converts the noncontainment condition d(R_H,R_R)>0 into a single linear inequality on u0. This makes the MPC a QP and gives recursive safety assurances without Hamilton-Jacobi precomputation.

Load-bearing premise

The guarantee rests on the human reachable set being a true conservative envelope: if a person moves faster than the assumed 1.0 m/s, accelerates harder, or extends a limb beyond the fixed radii, the robot's safe set can contain the real human state and no formal guarantee remains.

What would settle it

Pick the linear MAV model and one human, fix a horizon, and solve the MPC with Eq. (18); then brute-force search all admissible control sequences to check whether any state sequence that satisfies the per-step noncontainment condition can still end in collision (e.g., because the required escape direction changes from step to step faster than the drone can turn). A single such counterexample would falsify the recursive-safety claim. An experimental alternative: run the visual-servoing task with a person walking at 1.5 m/s, exceeding the assumed bound; a collision would show the guarantee's dep

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Safe MAV navigation among moving humans becomes a real-time quadratic program with linear safety constraints, solvable at 40 Hz on embedded hardware.
  • The method removes the need for extensive reachability precomputation, unlike Hamilton-Jacobi approaches, while retaining formal safety guarantees under stated assumptions.
  • Because the safety constraint is embodiment-independent, the same formulation can be ported to other robot platforms with linearizable dynamics.
  • Allowing overlaps between reachable sets—instead of requiring them disjoint—yields less conservative paths than standard forward reachability, as shown by higher efficiency in experiments.
  • The constraint degrades gracefully: when the initial state is infeasible (e.g., after perception failure), it is softened and the objective drives recovery.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The key insight—safety can be enforced on the executed input alone, not on all future inputs—might transfer to other safety-critical control settings (e.g., manipulators or autonomous vehicles), where conservative trajectory tubes are currently the bottleneck.
  • The paper's assumption of fixed human velocity/acceleration bounds is the main practical limit; a learned reachable set driven by current activity (walking, reaching, falling) would be a natural testable extension to reduce conservatism while preserving the guarantee.
  • The approach implies a trade-off: the human reachable set must grow with prediction horizon, so long horizons become expensive; the hybrid complex/simplified switch at 0.2s is a pragmatic answer, but adaptive switching times could be explored.
  • One could test whether the recursive guarantee survives imperfect tracking by injecting synthetic latency and noise into the joint observations in simulation, since the paper's real-world latency handling inflates reachable sets proportionally to unobserved time.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper presents HumanMPC (called HumanHalo in the abstract), a model predictive control framework for 3D MAV navigation in human-populated environments. The method tracks the full 24-joint SMPL human body, forecasts future motion with MotionMixer, and proposes a reachability-based safety constraint that, unlike standard disjoint-reachable-set formulations, constrains only the first control input u0 while accounting for future control freedom over the whole horizon. The central claim is that enforcing non-containment of the MAV reachable set in the human reachable set for the executed initial input guarantees the absence of inevitable collisions and recursive safety. Experiments in simulation with AMASS motions and on a real quadrotor are reported, comparing against distance-constraint baselines, forward-reachability variants, and a 2D navigation baseline.

Significance. If the safety derivation were sound, the contribution would be significant: a real-time, linear MPC safety constraint that works with a full 3D human-body model, avoids HJ-reachability precomputation, and remains computationally cheap. The paper provides real-world validation on an embedded MAV, clear experimental comparisons, and a sensible integration of a data-driven human motion forecaster. However, the theoretical core is not currently defensible: the distance function in Eq. (9)-(10) does not encode the stated non-containment condition, and the key linear constraint in Eq. (17)-(18) does not follow from the reachable set defined in Eq. (13). The recursive safety guarantee is asserted rather than proven. Because these issues directly affect the paper's central claim, the manuscript cannot be accepted in its present form.

major comments (4)
  1. [Section IV-B, Eqs. (8)-(10) and (15)] The stated constraint h_i>0 with d defined in Eq. (9) is not equivalent to R_R(u0) ⊄ R_H. For any overlapping but non-contained sets, the second case of Eq. (9) yields distance 0, so h_i=0 and the constraint is violated. Thus Eq. (10) forbids all overlap, directly contradicting the text and Fig. 2, which state that overlaps are safe. Moreover, h_i is defined as d(R_H, R_R), so in the dangerous case R_R ⊂ R_H the function evaluates to 0 (not negative), failing to identify the hazard. The separation formulation in Eq. (15) with ε>0 is a different condition. The manuscript must specify exactly which inequality is implemented and reconcile it with Eq. (8).
  2. [Section IV-B, Eqs. (12)-(18)] The linear constraint does not follow from the reachable set defined in Eq. (13). R_R in Eq. (13) depends only on x0 and the control bounds u_min,u_max, not on the optimized input u0; its center is Φx0 + Γu_c. Eq. (17) replaces this center with Φ1x0 + Γ1u0, but Φ1 and Γ1 are never defined. If they are stacked blocks of Φ and Γ, the remaining block contributions and the center term Γ_rem u_c,rem are dropped, and the half-extent e_R in Eq. (14) still includes variation of u0 that is no longer free. If instead R_R(u0) is intended to be the reachable set with the first input frozen, that set is different from Eq. (13) and is not derived. In neither case is Eq. (18) the support-function condition for the stated reachable set.
  3. [Section IV-B, paragraph after Eq. (18)] The claim that Eq. (18) 'recursively assures feasibility and therefore safety' is not proved. Non-containment at each k can be certified by different future input sequences; the paper does not show the existence of a single continuation, an invariant set, or a backwards reachable set argument, nor does it prove that the QP remains feasible at the next replan. The listed provisos (no saturation, accurate joints, identified model) do not replace the missing inductive step. A formal proof or a counterexample is needed.
  4. [Section V-A, Table II and AMASS validation] The human reachable-set bounds (v_i,max, a_i,max, ρ_i) are validated by comparing the reachable-set expansion against AMASS, and the simulation test scenarios are also drawn from AMASS. This makes the empirical validation circular for the core assumption that R_H is a true over-approximation of human motion. An out-of-distribution test set or a formal biomechanical justification of the bounds is required to support the claim that the safety guarantee holds beyond the specific motions used for calibration.
minor comments (5)
  1. [Title/Abstract] The abstract and arXiv metadata use the name 'HumanHalo', while the full paper uses 'HumanMPC' throughout. Please unify the name.
  2. [Table II] The table lists 'τmin −15° τmax 15°' twice; the second entry is presumably for θ or φ control limits. Please correct the label.
  3. [Section III-B and IV-B] Eq. (5) uses x for the stacked state vector, but Eq. (17) writes x_{k+1} = Φ1 x0 + Γ1 u0 with no definition of Φ1, Γ1, or the relationship to the stacked dynamics. This notation needs to be made precise.
  4. [Table I] The baseline row labeled 'None' omits the safety constraint entirely; consider clarifying in the baselines paragraph that this is the unconstrained MPC baseline, since the connection is easy to miss.
  5. [Section V-B] The real-world result 'never closer than 2.4 m' is compared to a safety margin of 0.5 m; please clarify whether 2.4 m already includes the MAV safety radius R_MAV and how the margin is defined in the real-world experiments.

Circularity Check

1 steps flagged

Partial empirical circularity: the human reachable-set parameters are validated on AMASS, and the simulated test obstacles are replayed from AMASS; the formal MPC safety derivation itself is not circular, though Eq. (18)'s recursive-safety claim is asserted rather than derived.

specific steps
  1. fitted input called prediction [Section V-A, Simulation Experiments (human reachable-set parameterization and test-set description)]
    "Although based on previous work in human motion analysis [31], [32], the parameterization of the reachable set is further validated by comparing its expansion with actual motions from the AMASS [34] dataset. ... Real human body motions are used as dynamic obstacles: 50 motion sequences are drawn from the AMASS [34] dataset, mostly focusing on walking scenarios, and replayed in simulation as the ground truth human motion."

    The human reachable-set parameters (radii, vi,max=1.0, ai,max=1.0 in Table II) are validated against AMASS motions, and the simulation safety evaluation then replays AMASS sequences as the ground-truth obstacles. If the reachable-set expansion was adjusted to cover AMASS, the simulated 100% collision-avoidance result is partly guaranteed by construction rather than independently tested. The formal safety theorem remains conditional on an over-approximating R_H, and the real-world trials provide independent evidence, but the AMASS simulation is not an independent validation of the conservativeness assumption.

full rationale

The core MPC safety formulation is a construction rather than a fitted prediction: Eq. (18) is a linear constraint imposed in the QP, and no target safety outcome is used to fit it. The main circularity concern is the AMASS overlap: the human reachable-set parameterization is 'validated' on AMASS and the simulation test obstacles are also drawn from AMASS, making the simulated safety results partly self-fulfilling. This is a partial empirical circularity, not a defect in the formal conditional claim. Self-citations such as [17], [19], and [35] are not load-bearing uniqueness theorems; they provide standard models/tools. Separately, the derivation from Eq. (13) to Eq. (18) has a real gap: R_R in Eq. (13) is u0-independent, while Eq. (17) substitutes a center depending on u0, so the claimed equivalence with the noncontainment condition is not established. That is a correctness/derivation issue, not circularity, and should be checked by reviewers; it does not raise the circularity score further.

Axiom & Free-Parameter Ledger

7 free parameters · 6 axioms · 0 invented entities

No new physical entities are introduced. The main extra degrees of freedom are hand-tuned or prior-work parameters that size the human reachable set and the safety margin; the recursive-safety guarantee also depends on unverified operational assumptions. These are not new forces, particles, or conserved quantities, so the invented-entity ledger is empty.

free parameters (7)
  • Human body-part radii ρHead, ρTorso, ρArm, ρHand = 0.2, 0.3, 0.205, 0.1 m
    Manually set initial radii for the spheres/capsules used to construct the human reachable set (Table II).
  • Human velocity and acceleration bounds v_i,max, a_i,max = 1.0, 1.0 (m/s, m/s^2)
    Conservative bounds used to inflate human reachable sets over the horizon; validated/compared against AMASS motions but not derived from a first-principles model.
  • Robot safety radius R_MAV = 0.5 m
    Added to human reachable-set dimensions to account for MAV size; hand-chosen and used as the collision threshold in experiments.
  • Reachability cost weight λ = 1000
    Weight of the reachability-cost term l_R^k in the MPC objective (Table II).
  • Reachable-set switching time t_s = 0.2 s
    Hand-chosen time after which the complex human reachable set switches to the simplified cylinder.
  • MAV linear model parameters c_i, b1..b4 = c_i=0.01, b1=1.0, b2=0.1, b3=1.0, b4=0.1
    System-identified parameters in the linearized MAV model (Eq. 3); taken from prior work/system identification rather than fitted to the safety outcome.
  • Separation margin ε = not specified
    Strictly positive margin in Eq. (15) needed for the supporting-hyperplane non-containment condition; no tuning value or selection procedure is given.
axioms (6)
  • standard math Convex separation: non-containment of one convex set in another is equivalent to existence of a supporting hyperplane separating at least part of the set (Eq. 15).
    Used to convert the geometric non-containment condition into a linear inequality; valid for convex sets.
  • domain assumption The MAV dynamics are well approximated by the linear second-order model (Eq. 3) under small attitude angles, vehicle symmetry, and negligible aerodynamic coupling.
    This linearization is asserted from [17],[19] and is used to claim the MAV reachable set is computed exactly.
  • domain assumption Human body motion is conservatively bounded by a double-integrator model with fixed radii, v_max, a_max, and inflated arm span.
    The human reachable-set construction (Sec. IV-B, Table II) relies on these bounds; if any real human motion exceeds them, safety is not guaranteed.
  • domain assumption Recursive feasibility conditions hold: the system starts in a feasible region, never operates at control bounds, has sufficiently accurate human-joint observations, and has a well-identified model.
    Stated immediately after Eq. (18) as the precondition for 'recursively assures feasibility and therefore safety'.
  • domain assumption The environment is free of non-human obstacles and occlusions.
    Acknowledged in the conclusion as a limitation: the framework does not account for occlusion or non-human obstacles.
  • domain assumption Perception error from back-projecting 2D pose detections onto the depth map is covered by inflating joint-wise reachable sets proportionally to unobserved time.
    The real-world pipeline assumes this inflation correctly captures observation uncertainty and latency (Sec. V-B).

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of HumanHalo -- Safe and Efficient 3D Navigation Among Humans via Minimally Conservative MPC." pith.science (2026). https://pith.science/paper/O34MIQ6Z

@misc{pith2026251017525,
  author       = {Pith},
  title        = {Pith review of: HumanHalo -- Safe and Efficient 3D Navigation Among Humans via Minimally Conservative MPC},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O34MIQ6Z}},
  note         = {Machine review of arXiv:2510.17525}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Safe and efficient robotic navigation among humans is essential for integrating robots into everyday environments. Most existing approaches focus on simplified 2D crowd navigation and fail to account for the full complexity of human body dynamics beyond root motion. We present HumanHalo, a Model Predictive Control (MPC) framework for 3D Micro Air Vehicle (MAV) navigation among humans that combines theoretical safety guarantees with data-driven models for realistic human motion forecasting. Our approach introduces a novel twist to reachability-based safety formulation that constrains only the initial control input for safety while modeling its effects over the entire planning horizon, enabling safe yet efficient navigation. We validate HumanHalo in both simulated experiments using real human trajectories and in the real-world, demonstrating its effectiveness across tasks ranging from goal-directed navigation to visual servoing for human tracking. While we apply our method to MAVs in this work, it is generic and can be adapted by other platforms. Our results show that the method ensures safety without excessive conservatism and outperforms baseline approaches in both efficiency and reliability.

Figures

Figures reproduced from arXiv: 2510.17525 by Helen Oleynikova, Sandra Hirche, Simon Schaefer, Stefan Leutenegger.

Figure 1
Figure 1. Figure 1: HumanMPC: A control framework for safe and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Based on the current MAV state x0 and the tracked 3D positions of human body joints, we first compute both the MAV’s and the human’s reachable sets over the entire horizon. We then optimize the control inputs uk with the requirement that the initial control input u0 must not lead to a situation where the MAV’s reachable set becomes a subset of the human’s reachable set at any time. While overlaps between t… view at source ↗
Figure 3
Figure 3. Figure 3: Example evaluation of our safety constraint for three [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Example construction of the MAV reachable set [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

36 extracted references · 3 linked inside Pith

  1. [1]

    Hierarchical drone navigation control for ensuring safety with stationary humans,

    R. Villalobos-Salazar, A. Contreras-Carlos, C. A. Toro-Arcila, and A. Morales-Diaz, “Hierarchical drone navigation control for ensuring safety with stationary humans,” in2024 21st International Conference on Electrical Engineering, Computing Science and Automatic Control (CCE), 2024, pp. 1–6

  2. [2]

    Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation,

    P. Trautman, J. Ma, R. M. Murray, and A. Krause, “Robot navigation in dense human crowds: Statistical models and experimental studies of human–robot cooperation,”The International Journal of Robotics Research, vol. 34, no. 3, pp. 335–356, 2015

  3. [3]

    Move beyond tra- jectories: Distribution space coupling for crowd navigation,

    M. Sun, F. Baldini, P. Trautman, and T. Murphey, “Move beyond tra- jectories: Distribution space coupling for crowd navigation,”Robotics: Science and Systems, 2021

  4. [4]

    Intention-aware online pomdp planning for autonomous driving in a crowd,

    H. Bai, S. Cai, N. Ye, D. Hsu, and W. S. Lee, “Intention-aware online pomdp planning for autonomous driving in a crowd,” in2015 IEEE International Conference on Robotics and Automation (ICRA), 2015, pp. 454–460

  5. [5]

    Robot planning with mathematical models of human state and action,

    A. D. Dragan, “Robot planning with mathematical models of human state and action,” 2017. [Online]. Available: https://arxiv.org/abs/1705.04226

  6. [6]

    Sicnav: Safe and interactive crowd navigation using model predictive control and bilevel optimization,

    S. Samavi, J. R. Han, F. Shkurti, and A. P. Schoellig, “Sicnav: Safe and interactive crowd navigation using model predictive control and bilevel optimization,”IEEE Transactions on Robotics, vol. 41, p. 801–818, 2025

  7. [7]

    End-to-end safe reinforcement learning through barrier functions for safety-critical continuous control tasks,

    R. Cheng, G. Orosz, R. M. Murray, and J. W. Burdick, “End-to-end safe reinforcement learning through barrier functions for safety-critical continuous control tasks,” 2019

  8. [8]

    Collision avoidance in pedestrian-rich environments with deep reinforcement learning,

    M. Everett, Y . F. Chen, and J. P. How, “Collision avoidance in pedestrian-rich environments with deep reinforcement learning,”IEEE Access, vol. 9, p. 10357–10377, 2021

  9. [9]

    Socially compliant mobile robot navigation via inverse reinforcement learning,

    H. Kretzschmar, M. Spies, C. Sprunk, and W. Burgard, “Socially compliant mobile robot navigation via inverse reinforcement learning,” The International Journal of Robotics Research, vol. 35, no. 11, pp. 1289–1307, 2016

  10. [10]

    A time-dependent hamilton- jacobi formulation of reachable sets for continuous dynamic games,

    I. Mitchell, A. Bayen, and C. Tomlin, “A time-dependent hamilton- jacobi formulation of reachable sets for continuous dynamic games,” IEEE Transactions on Automatic Control, vol. 50, no. 7, pp. 947–957, 2005

  11. [11]

    Planning, fast and slow: A framework for adaptive real-time safe trajectory planning,

    D. Fridovich-Keil, S. L. Herbert, J. F. Fisac, S. Deglurkar, and C. J. Tomlin, “Planning, fast and slow: A framework for adaptive real-time safe trajectory planning,”IEEE International Conference on Robotics and Automation, 2018

  12. [12]

    Reciprocal n- body collision avoidance,

    J. van den Berg, S. J. Guy, M. Lin, and D. Manocha, “Reciprocal n- body collision avoidance,” inRobotics Research, C. Pradalier, R. Sieg- wart, and G. Hirzinger, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 2011, pp. 3–19

  13. [13]

    Sicnav-diffusion: Safe and interactive crowd navigation with diffusion trajectory predictions,

    S. Samavi, A. Lem, F. Sato, S. Chen, Q. Gu, K. Yano, A. P. Schoellig, and F. Shkurti, “Sicnav-diffusion: Safe and interactive crowd navigation with diffusion trajectory predictions,”IEEE Robotics and Automation Letters, vol. 10, no. 9, p. 8738–8745, Sep. 2025

  14. [14]

    Leveraging neural network gradients within trajectory optimization for proactive human- robot interactions,

    S. Schaefer, K. Leung, B. Ivanovic, and M. Pavone, “Leveraging neural network gradients within trajectory optimization for proactive human- robot interactions,”IEEE International Conference on Robotics and Automation, 2020

  15. [15]

    Aircaprl: Autonomous aerial human motion capture using deep reinforcement learning,

    R. Tallamraju, N. Saini, E. Bonetto, M. Pabst, Y . T. Liu, M. Black, and A. Ahmad, “Aircaprl: Autonomous aerial human motion capture using deep reinforcement learning,”IEEE Robotics and Automation Letters, vol. 5, no. 4, pp. 6678–6685, Oct. 2020. [Online]. Available: https://ieeexplore.ieee.org/document/9158379

  16. [16]

    SMPL: A skinned multi-person linear model,

    M. Loper, N. Mahmood, J. Romero, G. Pons-Moll, and M. J. Black, “SMPL: A skinned multi-person linear model,”ACM Trans. Graphics (Proc. SIGGRAPH Asia), vol. 34, no. 6, pp. 248:1–248:16, Oct. 2015

  17. [17]

    Fully autonomous micro air vehicle flight and landing on a moving target using visual–inertial estimation and model-predictive control,

    D. Tzoumanikas, W. Li, M. Grimm, K. Zhang, M. Kovac, and S. Leutenegger, “Fully autonomous micro air vehicle flight and landing on a moving target using visual–inertial estimation and model-predictive control,”Journal of Field Robotics, vol. 36, no. 1, pp. 49–77, 2019. [Online]. Available: https://onlinelibrary.wiley.com/doi/abs/10.1002/rob.21821

  18. [18]

    Hybrid predictive control for aerial robotic physical interaction towards in- spection operations,

    G. Darivianakis, K. Alexis, M. Burri, and R. Siegwart, “Hybrid predictive control for aerial robotic physical interaction towards in- spection operations,” inIEEE International Conference on Robotics and Automation (ICRA), 2014, pp. 53–58

  19. [19]

    PAMPC: Perception- aware model predictive control for quadrotors,

    D. Falanga, P. Foehn, P. Lu, and D. Scaramuzza, “PAMPC: Perception- aware model predictive control for quadrotors,” inIEEE/RSJ Int. Conf. Intell. Robot. Syst. (IROS), 2018

  20. [20]

    End-to-end recovery of human shape and pose,

    A. Kanazawa, M. J. Black, D. W. Jacobs, and J. Malik, “End-to-end recovery of human shape and pose,” inComputer Vision and Pattern Recognition (CVPR), 2018

  21. [21]

    Vibe: Video inference for human body pose and shape estimation,

    M. Kocabas, N. Athanasiou, and M. J. Black, “Vibe: Video inference for human body pose and shape estimation,” inThe IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2020

  22. [22]

    Humans in 4D: Reconstructing and tracking humans with transform- ers,

    S. Goel, G. Pavlakos, J. Rajasegaran, A. Kanazawa*, and J. Malik*, “Humans in 4D: Reconstructing and tracking humans with transform- ers,” inInternational Conference on Computer Vision (ICCV), 2023

  23. [23]

    Poseformerv2: Exploring frequency domain for efficient and robust 3d human pose estimation,

    Q. Zhao, C. Zheng, M. Liu, P. Wang, and C. Chen, “Poseformerv2: Exploring frequency domain for efficient and robust 3d human pose estimation,” inConference on Computer Vision and Pattern Recogni- tion, 2023

  24. [24]

    Glopro: Globally- consistent uncertainty-aware 3d human pose estimation & tracking in the wild,

    S. Schaefer, D. F. Henning, and S. Leutenegger, “Glopro: Globally- consistent uncertainty-aware 3d human pose estimation & tracking in the wild,”IEEE International Conference on Intelligent Robots and Systems, 2023

  25. [25]

    Pace: Human and motion estimation from in-the-wild videos,

    M. Kocabas, Y . Yuan, P. Molchanov, Y . Guo, M. J. Black, O. Hilliges, J. Kautz, and U. Iqbal, “Pace: Human and motion estimation from in-the-wild videos,” in3DV, 2024

  26. [26]

    PARE: Part Attention Regressor for 3D Human Body Estimation,

    M. Kocabas, C.-H. P. Huang, O. Hilliges, and M. J. Black, “PARE: Part Attention Regressor for 3D Human Body Estimation,” inProceedings of the International Conference on Computer Vision (ICCV), 2021

  27. [27]

    Glamr: Global occlusion-aware human mesh recovery with dynamic cameras,

    Y . Yuan, U. Iqbal, P. Molchanov, K. Kitani, and J. Kautz, “Glamr: Global occlusion-aware human mesh recovery with dynamic cameras,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

  28. [28]

    Simple baselines for human pose estimation and tracking,

    B. Xiao, H. Wu, and Y . Wei, “Simple baselines for human pose estimation and tracking,”ECCV, 2018

  29. [29]

    Motionmixer: Mlp-based 3d human body pose forecasting,

    A. Bouazizi, A. Holzbock, U. Kressel, K. Dietmayer, and V . Bela- giannis, “Motionmixer: Mlp-based 3d human body pose forecasting,” inProceedings of the Thirty-First International Joint Conference on Artificial Intelligence, IJCAI-22. International Joint Conferences on Artificial Intelligence Organization, 7 2022, pp. 791–798

  30. [30]

    Boyd and L

    S. Boyd and L. Vandenberghe,Convex optimization. Cambridge university press, 2004

  31. [31]

    Sara: A tool for safe human-robot coexistence and collaboration through reachability analysis,

    S. R. Schepp, J. Thumm, S. B. Liu, and M. Althoff, “Sara: A tool for safe human-robot coexistence and collaboration through reachability analysis,” in2022 International Conference on Robotics and Automa- tion (ICRA), 2022, pp. 4312–4317

  32. [32]

    Pedestrian walking speed analysis: A systematic review,

    M. Giannoulaki and Z. Christoforou, “Pedestrian walking speed analysis: A systematic review,”Sustainability, vol. 16, no. 11, 2024. [Online]. Available: https://www.mdpi.com/2071-1050/16/11/4813

  33. [33]

    Hpipm: a high-performance quadratic programming framework for model predictive control,

    G. Frison and M. Diehl, “Hpipm: a high-performance quadratic programming framework for model predictive control,” 2020. [Online]. Available: https://arxiv.org/abs/2003.02547

  34. [34]

    AMASS: Archive of motion capture as surface shapes,

    N. Mahmood, N. Ghorbani, N. F. Troje, G. Pons-Moll, and M. J. Black, “AMASS: Archive of motion capture as surface shapes,” in International Conference on Computer Vision, Oct. 2019, pp. 5442– 5451

  35. [35]

    Okvis2: Realtime scalable visual-inertial slam with loop closure,

    S. Leutenegger, “Okvis2: Realtime scalable visual-inertial slam with loop closure,” 2022. [Online]. Available: https://arxiv.org/abs/2202.09199

  36. [36]

    Px4: A node-based multithreaded open source robotics framework for deeply embedded platforms,

    L. Meier, D. Honegger, and M. Pollefeys, “Px4: A node-based multithreaded open source robotics framework for deeply embedded platforms,” in2015 IEEE International Conference on Robotics and Automation (ICRA), 2015, pp. 6235–6240

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.