pith. machine review for the scientific record. sign in

arxiv: 2604.16670 · v1 · submitted 2026-04-17 · 💻 cs.RO

Recognition: unknown

Diffusion-Based Optimization for Accelerated Convergence of Redundant Dual-Arm Minimum Time Problems

Jushan Chen , Jonathan Fried , Santiago Paternain
Authors on Pith no claims yet

Pith reviewed 2026-05-10 08:00 UTC · model grok-4.3

classification 💻 cs.RO
keywords dual-arm robotsminimum-time optimizationdiffusion-based optimizationredundant manipulatorsCartesian path followingnonconvex optimizationprobabilistic sampling
0
0 comments X

The pith

Diffusion-based probabilistic sampling solves the nonconvex high-level problem for redundant dual-arm minimum-time path planning, cutting runtime by 35 times and Cartesian error by 34 percent.

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

The paper develops a framework that applies a new variant of model-based diffusion to minimize the time a redundant dual-arm robot needs to track a desired relative Cartesian path. Earlier bi-level methods solved the lower convex subproblem analytically but relied on gradient-based primal-dual updates for the upper nonconvex layer, creating high computation cost and preventing direct enforcement of an L-infinity error bound. Replacing those updates with probabilistic sampling from the diffusion model removes the gradient sparsity issue and accelerates the search. A reader would care because many dual-arm tasks in manufacturing or service robotics are limited by planning speed and by how tightly the end-effectors can stay on the intended path.

Core claim

The authors replace the gradient-based solver for the high-level nonconvex optimization with a novel variant of the model-based diffusion algorithm. Probabilistic sampling from this diffusion process generates candidate solutions for the joint-space trajectory parameters while the lower-level convex subproblem continues to be solved analytically; the combined procedure directly enforces an L-infinity Cartesian error constraint along the entire path and produces minimum-time solutions.

What carries the argument

A novel variant of the model-based diffusion algorithm that uses probabilistic sampling to explore and select high-quality solutions for the high-level nonconvex optimization problem.

If this is right

  • An L-infinity Cartesian error constraint can be imposed directly along the full joint trajectory without gradient sparsity problems.
  • High-level optimization runtime drops by a factor of 35 relative to the primal-dual baseline.
  • Cartesian tracking error of the resulting trajectories is 34 percent lower than the error obtained with the earlier gradient method.
  • The analytic solution of the lower-level convex subproblem combines directly with the sampled high-level decisions.

Where Pith is reading between the lines

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

  • The same sampling strategy could be tested on other bi-level robotic problems whose upper layer is nonconvex and whose gradients are sparse.
  • Because per-query time is greatly reduced, the method might support repeated re-planning inside a receding-horizon controller for changing task goals.
  • Hardware experiments on physical dual-arm platforms would reveal whether the simulated speed and accuracy gains survive actuator dynamics and sensor noise.

Load-bearing premise

The probabilistic sampling step consistently returns feasible, high-quality trajectories that respect the L-infinity Cartesian error bound without missing substantially better solutions.

What would settle it

Run the diffusion optimizer on a standard redundant dual-arm path-following benchmark, then check whether any generated trajectory violates the L-infinity Cartesian error limit at any sample point or whether the measured wall-clock time exceeds one-thirty-fifth of the prior gradient-based runtime.

Figures

Figures reproduced from arXiv: 2604.16670 by Jonathan Fried, Jushan Chen, Santiago Paternain.

Figure 1
Figure 1. Figure 1: Setup example consisting of two arms with 6 degrees of [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
read the original abstract

We present a framework leveraging a novel variant of the model-based diffusion algorithm to minimize the time required for a redundant dual-arm robot configuration to follow a desired relative Cartesian path. Our prior work proposed a bi-level optimization approach for the dual-arm problem, where we derived the analytical solution to the lower-level convex sub-problem and solved the high-level nonconvex problem using a primal-dual approach. However, the gradient-based nature leads to a large computation overhead, and it prohibits directly imposing an $L_{\infty}$ Cartesian error constraint along the joint trajectory due to the sparsity of the gradient. In this work, we propose a diffusion-based framework that relies on probabilistic sampling to tackle the aforementioned challenges in the nonconvex high-level problem, leading to a 35x reduction in the runtime and 34\% less Cartesian error compared to our prior work.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a novel variant of model-based diffusion for solving the high-level nonconvex subproblem in a bi-level formulation of minimum-time trajectory planning for redundant dual-arm robots tracking a relative Cartesian path. It contrasts this with the authors' prior primal-dual method, which suffers from high computational cost and inability to directly enforce an L∞ Cartesian error bound due to gradient sparsity, and reports a 35× runtime reduction together with 34% lower Cartesian error.

Significance. If the diffusion sampling reliably produces feasible, high-quality trajectories, the approach could materially accelerate real-time planning for redundant manipulators by replacing gradient-based nonconvex optimization with probabilistic sampling. The claimed speed-up is large enough to matter for online dual-arm coordination, but the absence of external baselines or shipped code makes it difficult to judge whether the gains generalize beyond the authors' own test cases.

major comments (2)
  1. [diffusion sampling procedure (around the description of the novel variant)] The central performance claims (35× runtime, 34% error reduction) rest on the diffusion variant generating trajectories that satisfy the hard L∞ Cartesian error constraint while escaping poor local minima. No section describes a concrete enforcement mechanism (hard projection, rejection sampling, or proven Lagrangian penalty) nor reports violation rates or feasibility statistics across sampled trajectories; without this evidence the constraint-satisfaction guarantee remains unverified.
  2. [experimental results and comparison tables] All numerical comparisons are performed exclusively against the authors' own prior bi-level primal-dual solver on the same test cases. The manuscript therefore provides no external benchmark (e.g., other nonconvex trajectory optimizers or sampling-based planners) that would establish whether the diffusion method improves upon the current state of the art rather than merely outperforming one specific earlier implementation.
minor comments (2)
  1. [methodology] Notation for the diffusion process (e.g., the precise form of the learned denoising network and the conditioning on the relative Cartesian path) should be stated more explicitly, ideally with a short pseudocode block, to allow reproduction.
  2. [abstract and §1] The abstract and introduction repeat the 35× and 34% figures without indicating the number of trials, standard deviation, or whether the numbers are means or best-case; adding these statistics would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below, indicating the revisions we will make to strengthen the paper.

read point-by-point responses

  1. Referee: The central performance claims (35× runtime, 34% error reduction) rest on the diffusion variant generating trajectories that satisfy the hard L∞ Cartesian error constraint while escaping poor local minima. No section describes a concrete enforcement mechanism (hard projection, rejection sampling, or proven Lagrangian penalty) nor reports violation rates or feasibility statistics across sampled trajectories; without this evidence the constraint-satisfaction guarantee remains unverified.

    Authors: We agree that explicit details on constraint enforcement are necessary to support the performance claims. The diffusion sampling procedure in our framework incorporates a hard projection step after each denoising iteration to enforce the L∞ Cartesian error bound, combined with rejection of any trajectories that remain infeasible after projection. In the revised manuscript, we will add a dedicated subsection under the diffusion sampling procedure that formally describes this mechanism, provides pseudocode for the full sampling loop, and reports empirical feasibility statistics including average rejection rates and the distribution of samples needed to obtain valid trajectories across all evaluated test cases. revision: yes

  2. Referee: All numerical comparisons are performed exclusively against the authors' own prior bi-level primal-dual solver on the same test cases. The manuscript therefore provides no external benchmark (e.g., other nonconvex trajectory optimizers or sampling-based planners) that would establish whether the diffusion method improves upon the current state of the art rather than merely outperforming one specific earlier implementation.

    Authors: We acknowledge that external benchmarks would help situate the results more broadly. Direct quantitative comparisons are difficult because few existing planners handle the exact bi-level formulation with a relative Cartesian path and hard L∞ error constraint. In the revision we will add a new discussion subsection that qualitatively contrasts our approach with representative methods such as RRT*-based planners and gradient-based nonconvex optimizers (e.g., CHOMP and IPOPT with random restarts), explaining the formulation mismatches that preclude straightforward numerical comparison. We will also include a limited additional experiment using a standard nonconvex solver on a subset of the test cases to provide an external reference point. revision: partial

Circularity Check

0 steps flagged

No significant circularity; new diffusion framework is independent of prior bi-level baseline.

full rationale

The paper introduces a diffusion-based optimization method as a distinct approach to the nonconvex high-level problem, explicitly contrasting it with the authors' earlier bi-level primal-dual method (cited only to highlight its gradient sparsity and runtime limitations). Performance numbers (35x runtime, 34% error reduction) are empirical outcomes of the new sampling procedure rather than quantities derived by construction from the prior work. No equations, uniqueness claims, or ansatzes reduce to self-citations or fitted inputs; the derivation chain for the model-based diffusion variant stands on its own probabilistic sampling mechanism. This is the common case of an incremental empirical improvement over self-authored baselines without circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The framework rests on the assumption that a model-based diffusion process can be adapted to produce feasible joint trajectories satisfying the relative Cartesian path constraint; no explicit free parameters, axioms, or invented entities are stated in the abstract.

pith-pipeline@v0.9.0 · 5443 in / 1144 out tokens · 28117 ms · 2026-05-10T08:00:08.141261+00:00 · methodology

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

55 extracted references · 4 canonical work pages · 2 internal anchors

  1. [1]

    Automatic control system for reach-to-grasp movement of a 7-dof robotic arm using ob- ject pose estimation with an rgb camera,

    S. Bai, J. Guo, Y . Jiang, H. Yokoi, and S. Togo, “Automatic control system for reach-to-grasp movement of a 7-dof robotic arm using ob- ject pose estimation with an rgb camera,” in2023 IEEE International Conference on Robotics and Biomimetics (ROBIO), pp. 1–6, 2023

    2023

  2. [2]

    Redundancy parameterization and inverse kinematics of 7-dof revolute manipulators,

    A. J. Elias and J. T. Wen, “Redundancy parameterization and inverse kinematics of 7-dof revolute manipulators,”Mechanism and Machine Theory, vol. 204, p. 105824, 2024

    2024

  3. [3]

    Clarifying the definition of redundancy as used in robotics,

    E. S. Conkur and R. Buckingham, “Clarifying the definition of redundancy as used in robotics,”Robotica, vol. 15, p. 583–586, 1997

    1997

  4. [4]

    Off-line programming of six-axis robots for optimum five-dimensional tasks,

    J. L ´eger and J. Angeles, “Off-line programming of six-axis robots for optimum five-dimensional tasks,”Mechanism and Machine Theory, vol. 100, pp. 155–169, 2016

    2016

  5. [5]

    Research on motion algorithm for intelligent operation of rock drill with eight degrees of freedom,

    Y . Xu, S. Tong, Z. Ju, Z. Wang, Y . Zhao, J. Yao, and Y . Zhao, “Research on motion algorithm for intelligent operation of rock drill with eight degrees of freedom,” in2018 IEEE ISR, pp. 61–65, 2018

    2018

  6. [6]

    Robotic deep rolling with iterative learning motion and force control,

    S. Chen, Z. Wang, A. Chakraborty, M. Klecka, G. Saunders, and J. Wen, “Robotic deep rolling with iterative learning motion and force control,”IEEE Robotics and Automation Letters, vol. 5, no. 4, pp. 5581–5588, 2020

    2020

  7. [7]

    iwaam: An automated system for monitoring and control of wire-arc additive manufacturing,

    F. Coutinho, N. Lizarralde, M. Mendes, R. Bostrom, T. Silva, M. Couto, and F. Lizarralde, “iwaam: An automated system for monitoring and control of wire-arc additive manufacturing,”IF AC- PapersOnLine, vol. 56, no. 2, pp. 6576–6581, 2023. 22nd IFAC World Congress

    2023

  8. [8]

    Kinematic control of redundant robot manipulators: A tutorial,

    B. Siciliano, “Kinematic control of redundant robot manipulators: A tutorial,”Journal of Intelligent and Robotic Systems, vol. 3, pp. 201– 212, 09 1990

    1990

  9. [9]

    Path tracking and obstacle avoidance for redundant robotic arms using fuzzy nmpc,

    A. M. Jasour and M. Farrokhi, “Path tracking and obstacle avoidance for redundant robotic arms using fuzzy nmpc,” in2009 American Control Conference, pp. 1353–1358, 2009

    2009

  10. [10]

    Finite-time model predictive tracking control of position and orientation for redundant manipulators,

    L. Jin, F. Zhang, M. Liu, and S. S.-D. Xu, “Finite-time model predictive tracking control of position and orientation for redundant manipulators,”IEEE Trans. on Industrial Electronics, vol. 70, no. 6, pp. 6017–6026, 2023

    2023

  11. [11]

    Error-summation enhanced newton algorithm for model predictive control of redundant manipulators,

    F. Zhang, L. Jin, and X. Luo, “Error-summation enhanced newton algorithm for model predictive control of redundant manipulators,” IEEE Trans. on Ind. Electronics, vol. 70, no. 3, pp. 2800–2811, 2023

    2023

  12. [12]

    A fault-tolerant method for motion planning of industrial redundant manipulator,

    Z. Li, C. Li, S. Li, and X. Cao, “A fault-tolerant method for motion planning of industrial redundant manipulator,”IEEE Transactions on Industrial Informatics, vol. 16, no. 12, pp. 7469–7478, 2020

    2020

  13. [13]

    Redundancy resolution in robots using parameterization through null space,

    Z. Kemeny, “Redundancy resolution in robots using parameterization through null space,”IEEE Transactions on Industrial Electronics, vol. 50, no. 4, pp. 777–783, 2003

    2003

  14. [14]

    Practical smooth minimum time trajectory planning for path following robotic manipulators,

    Q. Zhang, S.-R. Li, and X.-S. Gao, “Practical smooth minimum time trajectory planning for path following robotic manipulators,” in2013 American Control Conference, pp. 2778–2783, 2013

    2013

  15. [15]

    Minimum-time trajectory planning of a dual-manipulator ct system with synchroniza- tion and task constraints,

    H. Wang, M. Yuan, J. Huang, X. Zhang, and F. Liu, “Minimum-time trajectory planning of a dual-manipulator ct system with synchroniza- tion and task constraints,”IEEE Robotics and Automation Letters, vol. 10, no. 8, pp. 8475–8482, 2025

    2025

  16. [16]

    Simulation and trajectory generation of dual-robot collaborative welding for intersecting pipes,

    J. Xiong and Z. Fu, “Simulation and trajectory generation of dual-robot collaborative welding for intersecting pipes,”International Journal of Advanced Manufacturing Technology, vol. 111, 12 2020

    2020

  17. [17]

    Effects of traverse scanning speed of spray nozzle on the microstructure and mechanical properties of cold-sprayed ti6al4v coatings,

    A. W.-Y . Tan, W. Sun, Y . P. Phang, M. Dai, I. Marinescu, Z. Dong, and E. Liu, “Effects of traverse scanning speed of spray nozzle on the microstructure and mechanical properties of cold-sprayed ti6al4v coatings,”Journal of Thermal Spray Technology, vol. 26, pp. 1484– 1497, Oct 2017

    2017

  18. [18]

    The effect of traverse speed on deposition efficiency of cold sprayed stellite 21,

    R. A. Seraj, A. Abdollah-zadeh, S. Dosta, H. Canales, H. Assadi, and I. G. Cano, “The effect of traverse speed on deposition efficiency of cold sprayed stellite 21,”Surface and Coatings Technology, vol. 366, pp. 24–34, 2019

    2019

  19. [19]

    High-speed high-accuracy spatial curve tracking using motion primitives in industrial robots,

    H. He, C. lung Lu, Y . Wen, G. Saunders, P. Yang, J. Schoonover, A. Julius, and J. T. Wen, “High-speed high-accuracy spatial curve tracking using motion primitives in industrial robots,” 2023

    2023

  20. [20]

    Inverse kinematics in minimum-time trajectory planning for kinematically redundant ma- nipulators,

    A. Reiter, A. M ¨uller, and H. Gattringer, “Inverse kinematics in minimum-time trajectory planning for kinematically redundant ma- nipulators,” inIECON 2016 - 42nd Annual Conference of the IEEE Industrial Electronics Society, pp. 6873–6878, 2016

    2016

  21. [21]

    Resolved motion rate control of manipulators and human prostheses,

    D. E. Whitney, “Resolved motion rate control of manipulators and human prostheses,”IEEE Transactions on Man-Machine Systems, vol. 10, no. 2, pp. 47–53, 1969

    1969

  22. [22]

    Inverse kinematic functions for redundant manipulators,

    C. Wampler, “Inverse kinematic functions for redundant manipulators,” inProceedings. 1987 IEEE International Conference on Robotics and Automation, vol. 4, pp. 610–617, 1987

    1987

  23. [23]

    A trust region method based on interior point techniques for nonlinear programming,

    R. H. Byrd, J. C. Gilbert, and J. Nocedal, “A trust region method based on interior point techniques for nonlinear programming,”Math- ematical programming, vol. 89, pp. 149–185, 2000

    2000

  24. [24]

    An interior algorithm for nonlinear optimization that combines line search and trust region steps,

    R. A. Waltz, J. L. Morales, J. Nocedal, and D. Orban, “An interior algorithm for nonlinear optimization that combines line search and trust region steps,”Mathematical programming, vol. 107, no. 3, pp. 391–408, 2006

    2006

  25. [25]

    P. E. Gill, W. Murray, and M. H. Wright,Practical optimization. SIAM, 2019

    2019

  26. [26]

    CasADi – A software framework for nonlinear optimization and optimal control,

    J. A. E. Andersson, J. Gillis, G. Horn, J. B. Rawlings, and M. Diehl, “CasADi – A software framework for nonlinear optimization and optimal control,”Mathematical Programming Computation, 2018

    2018

  27. [27]

    Composite- learning-based adaptive neural control for dual-arm robots with rela- tive motion,

    Y . Jiang, Y . Wang, Z. Miao, J. Na, Z. Zhao, and C. Yang, “Composite- learning-based adaptive neural control for dual-arm robots with rela- tive motion,”IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 3, pp. 1010–1021, 2022

    2022

  28. [28]

    Redundancy resolution for dual-arm robots inspired by human asymmetric bimanual action: Formulation and experiments,

    J. Lee and P. H. Chang, “Redundancy resolution for dual-arm robots inspired by human asymmetric bimanual action: Formulation and experiments,” in2015 IEEE International Conference on Robotics and Automation (ICRA), pp. 6058–6065, 2015

    2015

  29. [29]

    Quadratic programming based inverse kinematics for precise bimanual manipulation,

    T. Chaki and T. Kawakami, “Quadratic programming based inverse kinematics for precise bimanual manipulation,” in2024 IEEE Inter- national Conference on Robotics and Automation (ICRA), pp. 16024– 16030, 2024

    2024

  30. [30]

    Deep learning- based optimization for motion planning of dual-arm assembly robots,

    K.-C. Ying, P. Pourhejazy, C.-Y . Cheng, and Z.-Y . Cai, “Deep learning- based optimization for motion planning of dual-arm assembly robots,” Computers & Industrial Engineering, vol. 160, p. 107603, 2021

    2021

  31. [31]

    Simultaneous dual-arm motion planning for minimizing operation time,

    J. Kurosu, A. Yorozu, and M. Takahashi, “Simultaneous dual-arm motion planning for minimizing operation time,”Applied Sciences, vol. 7, no. 12, 2017

    2017

  32. [32]

    Force and time-optimal trajectory planning for dual-arm unilateral cooperative grasping,

    R. Di Leva, H. Gattringer, A. M ¨uller, and M. Carricato, “Force and time-optimal trajectory planning for dual-arm unilateral cooperative grasping,”Mechanism and Machine Theory, vol. 201, p. 105729, 2024

    2024

  33. [33]

    A bi-level optimization method for re- dundant dual-arm minimum time problems,

    J. Fried and S. Paternain, “A bi-level optimization method for re- dundant dual-arm minimum time problems,”IEEE Control Systems Letters, vol. 9, pp. 1321–1326, 2025

    2025

  34. [34]

    Time-optimal trajectory generation for dynamic vehicles: A bilevel optimization approach,

    G. Tang, W. Sun, and K. Hauser, “Time-optimal trajectory generation for dynamic vehicles: A bilevel optimization approach,” in2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 7644–7650, 2019

    2019

  35. [35]

    Fast uav trajectory optimization us- ing bilevel optimization with analytical gradients,

    W. Sun, G. Tang, and K. Hauser, “Fast uav trajectory optimization us- ing bilevel optimization with analytical gradients,”IEEE Transactions on Robotics, vol. 37, no. 6, pp. 2010–2024, 2021

    2010

  36. [36]

    Joint trajectory optimization for redundant manipulators with constant path speed,

    J. Fried and S. Paternain, “Joint trajectory optimization for redundant manipulators with constant path speed,” in2024 American Control Conference (ACC), pp. 3833–3840, 2024

    2024

  37. [37]

    Denoising Diffusion Probabilistic Models

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,”ArXiv, vol. abs/2006.11239, 2020

    work page internal anchor Pith review arXiv 2006

  38. [38]

    Denoising Diffusion Implicit Models

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,”ArXiv, vol. abs/2010.02502, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  39. [39]

    Diffusion policy: Visuomotor policy learning via action diffusion,

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel, R. Tedrake, and S. Song, “Diffusion policy: Visuomotor policy learning via action diffusion,”The International Journal of Robotics Research, p. 02783649241273668, 2023

    2023

  40. [40]

    On-device diffusion transformer policy for efficient robot manipulation,

    Y . Wu, H. Wang, Z. Chen, J. Pang, and D. Xu, “On-device diffusion transformer policy for efficient robot manipulation,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2025

    2025

  41. [41]

    Safediffuser: Safe planning with diffusion probabilistic models,

    W. Xiao, T.-H. Wang, C. Gan, R. Hasani, M. Lechner, and D. Rus, “Safediffuser: Safe planning with diffusion probabilistic models,” in The Thirteenth International Conference on Learning Representations, 2023

    2023

  42. [42]

    Model-based diffusion for trajectory optimization,

    C. Pan, Z. Yi, G. Shi, and G. Qu, “Model-based diffusion for trajectory optimization,” 2024

    2024

  43. [43]

    Con- strained diffusers for safe planning and control,

    J. Zhang, L. Zhao, A. Papachristodoulou, and J. Umenberger, “Con- strained diffusers for safe planning and control,” 2025

    2025

  44. [44]

    Equality constrained diffusion for direct trajectory optimization,

    V . Kurtz and J. W. Burdick, “Equality constrained diffusion for direct trajectory optimization,” 2024

    2024

  45. [45]

    arXiv preprint arXiv:2505.21851 , year=

    S. Jiang, X. Fang, N. Roy, T. Lozano-P ´erez, L. P. Kaelbling, and S. An- cha, “Streaming flow policy: Simplifying diffusion/flow-matching policies by treating action trajectories as flow trajectories,”arXiv preprint arXiv:2505.21851, 2025

    work page arXiv 2025

  46. [46]

    Pad-tro: Projection-augmented diffusion for direct trajectory optimization,

    J. Chen and S. Paternain, “Pad-tro: Projection-augmented diffusion for direct trajectory optimization,”ArXiv, vol. abs/2510.04436, 2025

    work page arXiv 2025

  47. [47]

    Generative predictive control: Flow matching policies for dynamic and difficult-to-demonstrate tasks,

    V . Kurtz and J. W. Burdick, “Generative predictive control: Flow matching policies for dynamic and difficult-to-demonstrate tasks,” in Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2026. To appear

    2026

  48. [48]

    Diffusion model predictive control,

    G. S. Zhou, S. Swaminathan, R. V . Raju, J. S. Guntupalli, W. Lehrach, J. Ortiz, A. Dedieu, M. L ´azaro-Gredilla, and K. Murphy, “Diffusion model predictive control,” inProceedings of the International Confer- ence on Learning Representations (ICLR), 2026. Generative Models for Robot Learning workshop

    2026

  49. [49]

    A more compact expression of relative jacobian based on individual manipulator jacobians,

    R. S. Jamisola and R. G. Roberts, “A more compact expression of relative jacobian based on individual manipulator jacobians,”Robotics and Autonomous Systems, vol. 63, pp. 158–164, 2015

    2015

  50. [50]

    Optimization of parame- ters for curve interpolation by cubic splines,

    A. Kouibia, M. Pasadas, and M. Rodr ´ıguez, “Optimization of parame- ters for curve interpolation by cubic splines,”J. of Computational and Applied Mathematics, vol. 235, no. 14, pp. 4187–4198, 2011

    2011

  51. [51]

    Obstacle-avoidance path planning of unmanned vehicles based on polynomial optimization,

    S. Zhu, J. Yuan, J. Xiao, Y . Wang, Z. Pan, P. Zhao, and K. Feng, “Obstacle-avoidance path planning of unmanned vehicles based on polynomial optimization,” in2022 China Automation Congress (CAC), pp. 6988–6992, IEEE, 2022

    2022

  52. [52]

    Modeling the kinematics of human locomotion over continuously varying speeds and inclines,

    K. R. Embry, D. J. Villarreal, R. L. Macaluso, and R. D. Gregg, “Modeling the kinematics of human locomotion over continuously varying speeds and inclines,”IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 26, no. 12, pp. 2342–2350, 2018

    2018

  53. [53]

    Chapter 3 – the cross-entropy method for optimization,

    Z. I. Botev, D. P. Kroese, R. Y . Rubinstein, and P. L’Ecuyer, “Chapter 3 – the cross-entropy method for optimization,” inHandbook of Statistics (C. R. Rao and V . Govindaraju, eds.), vol. 31, pp. 35–59, Elsevier, Jan. 2013

    2013

  54. [54]

    Boyd and L

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

    2004

  55. [55]

    CasADi – A software framework for nonlinear optimization and optimal control,

    J. A. E. Andersson, J. Gillis, G. Horn, J. B. Rawlings, and M. Diehl, “CasADi – A software framework for nonlinear optimization and optimal control,”Mathematical Programming Computation, vol. 11, no. 1, pp. 1–36, 2019

    2019