Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

STDArm: Transferring Visuomotor Policies From Static Data Training to Dynamic Robot Manipulation

T0 review · 5 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read STDArm shows that static-trained visuomotor policies can transfer to moving robot platforms through real-time action correction, recovering near-static success rates without retraining.

desk verdict A practical wrapper for deploying static-trained policies on moving robots, with consistent gains across platforms, but the headline experiments rely on 15 trials with manually shaken bases and no controlled disturbance, so the quantitative claims are indicative rather than proven. read the letter →

arxiv 2504.18792 v1 pith:IJXBKVEU submitted 2025-04-26 cs.RO

Yifan Duan , Heng Li , Yilong Wu , Wenhao Yu , Xinran Zhang , Yedong Shen , Jianmin Ji , Yanyong Zhang This is my paper · ORCID
classification cs.RO
keywords STDArmvisuomotorpoliciesimitationlearningmobilemanipulationactioncorrectionposepredictionlatencycompensationdrone
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

STDArm is a system for moving imitation-learned visuomotor policies from fixed tabletops to moving platforms, including a drone. The paper's central claim is that a real-time action-correction layer, rather than retraining or recollecting demonstrations, is enough to recover near-static manipulation performance while the base moves. This matters because high-quality demonstrations are far easier to collect on a fixed arm, and existing mobile-manipulation pipelines either require specialized data collection or accept large drops in success. In experiments with diffusion policy and ACT on three tasks, STDArm raises dynamic success rates from near zero to values close to the static baselines. The system runs entirely on an edge computing board.

What carries the argument

The mechanism that carries the argument is the stabilizer's action-correction transform. At each high-frequency control step $\tau$, with total system latency $\Delta t$, extrinsic calibration $E$ between the SLAM frame and the arm frame, and platform pose stream at frequency $f$, the stabilizer forms $\delta(\tau,\Delta t,E) = E^{-1}\left(\Delta p_{\lceil -\Delta t \tau/f \rceil}^{-1}\,\Delta p^*_{\lfloor \Delta t/f \rfloor}\right)^{-1}E$, where $\Delta p$ is an observed past relative pose and $\Delta p^*$ a predicted future relative pose from the lightweight network. The final command is $A'_\tau = \delta(\tau,\Delta t,E)\, A_\tau$, so the arm's motion relative to the moving base is what the policy intended relative to the world. This correction works only because the action manager first provides a dense 50 Hz action stream by temporal ensembling overlapping action chunks with exponentially decaying weights $w_{t+i} = \exp(-\alpha i)$ and linearly interpolating between buffer entries; the prediction network itself is a parallel long short-term memory (LSTM) and gated recurrent unit (GRU) whose concatenated outputs feed a fully connected layer.

What would settle it

Record success with STDArm active while moving the base to viewpoints or lighting conditions not present in the static training data, for example rotating the camera to look at the object from above or swapping the background. If success drops to the uncorrected baseline level, the claim that STDArm preserves manipulation capability under platform motion is disproved for those conditions.

Watch

Extended reading notes

Core claim

On the paper's own terms, STDArm's discovery is that a static-trained visuomotor policy can be transplanted to a moving platform by treating deployment as a latency and frequency alignment problem rather than a policy-learning problem. The policy network is untouched; instead, an action manager raises and stabilizes the action stream, a stabilizer predicts the platform's short-term motion with a lightweight LSTM-GRU network and applies the correction $\delta(\tau,\Delta t,E)$ to every action, and a warm-up routine calibrates the total system latency online. The quantitative claim is that this restores near-static success: DP+STDArm on the mobile-table arm goes from 0% to 53.3% on Pick and Place and from 20% to 93.3% on Get the Delivery, and the drone version of Get the Delivery rises from 13.3% to 40%, while static performance is preserved.

Load-bearing premise

The visuomotor policy must still recognize the target from images; STDArm only corrects actions, so if the moving camera's changing viewpoint or background defeats the policy's perception, the system cannot restore the task.

Editorial extensions

If this is right

  • Static demonstration datasets, which are cheap to collect on fixed arms, become usable for drones, quadrupeds, and other moving bases without recollecting data on the moving platform.
  • Action-chunked visuomotor policies can be deployed at higher effective control frequency on low-cost edge hardware, improving responsiveness without changing the policy network.
  • Ablations on the leadscrew setup show the components are complementary: the full STDArm reaches 80% on Get the Delivery, while removing online latency estimation or the action manager drops success to 53.3% or below.
  • The drone result, though modest at 40% versus 13.3%, suggests aerial manipulation is within reach of static-trained policies once motion compensation and latency are handled.

Reading between the lines

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

  • An extension the paper suggests but does not demonstrate is that any action-chunked policy outputting end-effector poses could be wrapped by the same layer, since the policy network's weights are never modified.
  • Because the correction depends on a visual-inertial SLAM pose stream, the system will inherit SLAM drift or tracking loss in low-texture or outdoor environments; the paper does not evaluate these failure modes.
  • The pre-mission latency search could be replaced by continuous online calibration, which would track changing compute loads on shared edge devices; this is a testable improvement rather than a paper claim.
  • The LS-3DoF ablation shows stronger gains on regular, predictable motion, suggesting the predictor exploits predictability; isolating motion predictability as an independent variable would separate this effect from pure latency compensation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper presents STDArm, a modular system for transferring visuomotor policies trained under static conditions to mobile robotic platforms. The system consists of three components: an action manager that performs temporal ensembling and interpolation to raise control frequency, a prediction-network-based stabilizer that compensates for platform motion using high-frequency visual SLAM poses, and an online latency estimation module that calibrates system delay during a pre-mission warm-up. The authors evaluate STDArm on four platform configurations (mobile table, mobile frame, leadscrew, drone), two policies (Diffusion Policy and ACT), and three manipulation tasks, reporting that STDArm substantially improves success rates under dynamic conditions and preserves static performance. The core claim is that STDArm enables real-time motion compensation with centimeter-level precision without modifying the foundational policy.

Significance. The problem addressed is timely and practical: static-trained visuomotor policies are difficult to deploy on legged, wheeled, and aerial platforms because of platform motion, inference latency, and limited onboard compute. STDArm's design is appealing in its modularity: it does not require retraining the policy, works with multiple policy architectures, and runs entirely on edge hardware. If the reported gains are reproducible, the system would be a useful and low-cost tool for mobile manipulation. The experiments span multiple platforms, arms, and tasks, and the authors include ablations on a programmable leadscrew platform to control for some sources of variability. The paper also provides a project page and builds on public codebases. However, the central quantitative claim is currently supported only by a small number of trials under manually generated and unmeasured disturbances, which makes it difficult to attribute the observed improvements to STDArm specifically. The latency-estimation procedure also has a circularity issue in one of the demonstration tasks.

major comments (5)
  1. [Sec. V-A (Experimental Setup) and Table I] The headline dynamic results for MT-3DoF and MF-5DoF rely on manually shaking the platform, but the disturbance is neither recorded nor matched between the baseline and STDArm conditions. With only 15 trials per cell, a difference in shaking amplitude or frequency between runs could explain part of the reported gap (for example, Pick and Place on MT-3DoF from 0% to 53.3%). The double-blind protocol is commendable, but it does not guarantee that the baseline and STDArm trials face equivalent disturbances. The reproducible LS-3DoF platform is used only for ablations, not for the headline baseline-versus-STDArm comparison. A controlled or at least recorded-disturbance experiment is needed before the gains in Table I can be attributed to STDArm rather than to trial difficulty.
  2. [Table I and Fig. 7] Every reported success rate is based on 15 trials, with no error bars, confidence intervals, or significance tests. Several differences are within a few trials: for example, 66.7% vs. 53.3% in Stack Cylinder on MF-5DoF, and 40% vs. 13.3% on UAV-3DoF. Without statistical treatment or a larger number of trials, the claimed 'recovery' of static-level performance is not firmly established. At minimum, exact binomial confidence intervals should be reported, and the authors should state how many trials were performed per cell and whether trials were randomized in blocks.
  3. [Sec. III-D and Sec. V-D (Fig. 8)] The system latency is calibrated online by minimizing the red marker's pixel-plane movement in an end-hold task (Sec. III-D), and the same end-hold task and the same marker-movement metric are then used in Sec. V-D to demonstrate the stabilizer's benefit (Fig. 8). This is partly circular: the configuration is tuned to optimize exactly the metric used for evaluation. The demonstration would be more convincing if it used an independent metric, such as measured end-effector position error under held-out motion conditions, or if the end-hold task success was compared against a separate task objective.
  4. [Sec. V-A (UAV-3DoF) and Table I] For the UAV experiment, the text states that 'an operator manually maintains the relative positioning between the drone and the delivery target.' This means the 40% success rate with STDArm does not demonstrate autonomous compensation for platform drift; it shows only that STDArm can help under human-assisted station-keeping. The related claim in Sec. V-B that the UAV result highlights 'adaptability to ... inherent aerodynamic uncertainties' is therefore overstated. To support the aerial-platform claim, the experiment would need to quantify the platform's position error and separate autonomous stabilization from operator assistance.
  5. [Abstract and Sec. VI] The paper repeatedly claims 'centimeter-level operational precision' and 'centimeter-level precision,' but no direct precision measurement is reported. The tasks have tolerance specifications (e.g., 4 cm for stacking), but success rates are not equivalent to a precision measurement. The end-hold heatmaps in Fig. 8 are qualitative. To support this claim, the authors should report actual end-effector error statistics (mean, std, or a percent-within-threshold) under dynamic conditions.
minor comments (6)
  1. [Eq. (6)] The action interpolation formula appears to have its weights reversed: at τ = t it gives A_{t+1} instead of A_t. As written, A_τ = (τ−t)A_t + (t+1−τ)A_{t+1}; the intended formula is likely A_τ = (t+1−τ)A_t + (τ−t)A_{t+1}. Please correct the equation or clarify the definition of τ.
  2. [Fig. 7] The axis labels in Fig. 7 are garbled/unreadable in the provided version; the figure should be regenerated with clear, legible labels.
  3. [Sec. III-C, Eq. (9)] The definition of Δt_τ is not explicitly distinguished from Δt elsewhere in the text; please clarify that Δt_τ is the interval from action generation to execution for action Aτ, while Δt is the total latency estimated online.
  4. [Sec. V-A] The training of the pose prediction network is described only as 'approximately five minutes of motion data.' Please provide details on the supervision signal and how future poses are obtained for training, as this is central to the stabilizer's operation.
  5. [Intro, Sec. I] The abstract states the SLAM operates at '10× the base policy's observation rate.' With SLAM at 200 Hz and the DP policy at 5 Hz, the factor is 40×; for ACT at 30 Hz it is about 6.7×. Please make this statement consistent with the actual frequencies.
  6. [Throughout] There are several typographical artifacts, such as 'UA V-3DoF' with an extra space and inconsistent formatting in the platform names; these should be cleaned up.

Circularity Check

1 steps flagged · score 2.0 of 10

End-hold demonstration reuses the same marker-stability objective used to calibrate the latency; main manipulation benchmarks remain independent.

  1. fitted input called prediction [Sec. III-D (Latency Estimation) and Sec. V-D (Accuracy of End Hold), Fig. 8]
    "We conduct a linear search for the latency starting from zero, identifying the latency that minimizes the marker's movement as the system delay. ... To more intuitively demonstrate the contribution of our method to end-effector stability, we conduct an end-effector holding task on MT-3DoF to visualize performance. ... tracking the movement trajectory of a red ball held in the gripper."

    The system latency is calibrated by minimizing the red marker's positional variance in the end-hold task (Sec. III-D). The end-hold demonstration in Sec. V-D then evaluates the stabilizer using exactly the same red-marker movement criterion. The favorable STDArm heatmap is therefore partly a restatement of the calibration objective: the selected latency is, by construction, the value that minimizes that metric. This makes the end-hold figure a consistency check of the optimizer rather than an independent validation of the stabilizer. However, the main manipulation tasks (Pick and Place, Stack Cylinder, Get the Delivery) use external success criteria that were not optimized during latency calibration, so the central transfer claim retains independent content.

full rationale

STDArm's core derivation is not circular. The policy network is trained on static demonstrations; the pose prediction network is trained on separate motion data; Eq. (9)-(10) provide a kinematic compensation using the estimated latency; and the headline results are measured by task success rates, which are external to the calibration procedure. The action manager, temporal ensemble, and interpolation are evaluated by ablations on task success rather than by a self-defined metric. The only notable circular element is the end-hold visualization in Sec. V-D, whose marker-stability metric is exactly the objective minimized during latency estimation in Sec. III-D. That demonstration is illustrative, not load-bearing for the central claim. The manual-shaking protocol is a reproducibility and fairness concern, not a circularity concern, and the paper's self-citations are not load-bearing. Overall, the derivation is self-contained with one minor circularity in the supporting end-hold figure, so a score of 2 is appropriate.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new physical entities are postulated. The free parameters are calibration values and a hyperparameter; the axioms are assumptions about the action representation, SLAM stability, predictor generalization, and base-policy robustness.

free parameters (3)
  • Temporal ensemble decay α = 1
    Hyperparameter in Eq 4 controlling exponential weighting of new actions; chosen by hand (Sec IV-A), affects smoothing and responsiveness.
  • System latency Δt = not reported
    Estimated online by linear search minimizing marker motion in the warm-up routine (Sec III-D); used in Eq 9 for stabilization. It is a per-platform, per-condition calibration, effectively fitted to the end-hold criterion.
  • Extrinsics E between SLAM and arm frames = not reported
    Manually pre-calibrated transform used in Eq 9-10; central to correcting arm actions by platform motion.
assumptions (4)
  • domain assumption The policy action u is an end-effector position/pose expressed in a frame where the multiplicative correction δ in Eq 9 can be applied to compensate platform motion.
    Eq 10 A'τ = δ·Aτ assumes actions are rigid-body transforms in the arm's base frame and that the low-level arm controller can execute the corrected setpoint; no validation of this representation for the 3DoF and 5DoF arms.
  • domain assumption The visual SLAM pose estimates are accurate and the single estimated latency Δt captures the total pipeline delay throughout task execution.
    Sec III-C/III-D assume pose estimation and latency are stable; time-varying delays or pose drift would invalidate the compensation.
  • domain assumption The pose prediction network, trained on about five minutes of platform motion data, generalizes to the perturbations encountered at test time.
    Sec V-A trains the predictor on a short motion sequence; generalization to new shaking patterns or drone hover disturbances is assumed.
  • domain assumption The foundational policy network remains functional when the camera moves with the platform.
    Explicitly acknowledged in Limitations: changes in camera perspective and background during movement may break the policy; STDArm does not address this.

how reviews work

0 comments
Cite this review

Pith. "Pith review of STDArm: Transferring Visuomotor Policies From Static Data Training to Dynamic Robot Manipulation." pith.science (2026). https://pith.science/paper/IJXBKVEU

@misc{pith2026250418792,
  author       = {Pith},
  title        = {Pith review of: STDArm: Transferring Visuomotor Policies From Static Data Training to Dynamic Robot Manipulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IJXBKVEU}},
  note         = {Machine review of arXiv:2504.18792}
}
read the original abstract

Recent advances in mobile robotic platforms like quadruped robots and drones have spurred a demand for deploying visuomotor policies in increasingly dynamic environments. However, the collection of high-quality training data, the impact of platform motion and processing delays, and limited onboard computing resources pose significant barriers to existing solutions. In this work, we present STDArm, a system that directly transfers policies trained under static conditions to dynamic platforms without extensive modifications. The core of STDArm is a real-time action correction framework consisting of: (1) an action manager to boost control frequency and maintain temporal consistency, (2) a stabilizer with a lightweight prediction network to compensate for motion disturbances, and (3) an online latency estimation module for calibrating system parameters. In this way, STDArm achieves centimeter-level precision in mobile manipulation tasks. We conduct comprehensive evaluations of the proposed STDArm on two types of robotic arms, four types of mobile platforms, and three tasks. Experimental results indicate that the STDArm enables real-time compensation for platform motion disturbances while preserving the original policy's manipulation capabilities, achieving centimeter-level operational precision during robot motion.

Figures

Figures reproduced from arXiv: 2504.18792 by the authors.

Figure 1
Figure 1. When deploying a policy trained on static data on computationally [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Pipeline of STDArm. Given images and joint state observations, our system first generates a low-frequency action sequence via a policy network. These actions are then passed to action manager, which maintains an action buffer using temporal ensemble and generates high-frequency actions through interpolation. Subsequently, a stabilizer refines the action based on estimated latency and real-time pose predictions to co… view at source ↗
Figure 3
Figure 3. By utilizing the pose at action generation, the predicted pose, and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Four experimental configurations we design. Each is equipped with two cameras for observation inputs and a T265 camera for pose estimation. The [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: We design three tasks of varying difficulty levels and conduct experiments across three experimental configurations. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Temporal task progression during experiments. Beyond task completion metrics, the motion of the platforms also worth noting. Videos on the [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Ablation Study with MT-3DoF and MF-5DoF on Get the Delivery TABLE II ABLATION STUDY WITH LS-3DoF ON GET THE DELIVERY. No. Action Manager Latency=0 Latency Estimation Success Temporal Rate(%) Ensemble Action Interpolation 0 0 1 ✓ 20 2 ✓ 40 3 ✓ ✓ 26.6 4 ✓ ✓ 33.3 5 ✓ ✓ ✓ …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. CAC-VLA: Context-Gated Action Conditioning for Vision-Language-Action Models

    cs.RO 2026-07 conditional novelty 6.0 of 10

    VLM-predicted coarse-to-fine latent actions, injected through a learned context gate, improve continuous action-expert control on LIBERO (98.3%) and LIBERO-Plus (89.5%).

Reference graph

Works this paper leans on

37 extracted references · 14 canonical work pages · cited by 1 Pith paper

  1. [1]

    Deep generative models in robotics: A survey on learning from multimodal demonstrations

    Julen Urain, Ajay Mandlekar, Yilun Du, Mahi Shafiullah, Danfei Xu, Katerina Fragkiadaki, Georgia Chalvatzaki, and Jan Peters. Deep generative models in robotics: A survey on learning from multimodal demonstrations. arXiv preprint arXiv:2408.04380 , 2024

    arXiv 2024

  2. [2]

    Vision-based multi-task manipulation for inexpensive robots using end-to-end learning from demonstration

    Rouhollah Rahmatizadeh, Pooya Abolghasemi, Ladislau B¨ol¨oni, and Sergey Levine. Vision-based multi-task manipulation for inexpensive robots using end-to-end learning from demonstration. In 2018 IEEE international conference on robotics and automation (ICRA) , pages 3758–3765. IEEE, 2018

    work page 2018

  3. [3]

    Learning fine-grained bimanual manipulation with low-cost hardware

    Tony Z Zhao, Vikash Kumar, Sergey Levine, and Chelsea Finn. Learning fine-grained bimanual manipulation with low-cost hardware. Robotics: Science and Systems (RSS) , 2023

    work page 2023

  4. [4]

    Diffusion policy: Visuomotor policy learning via action diffusion

    Cheng Chi, Zhenjia Xu, Siyuan Feng, Eric Cousineau, Yilun Du, Benjamin Burchfiel, Russ Tedrake, and Shuran Song. Diffusion policy: Visuomotor policy learning via action diffusion. The International Journal of Robotics Research, page 02783649241273668, 2023

    2023

  5. [5]

    Fabricflownet: Bi- manual cloth manipulation with a flow-based policy

    Thomas Weng, Sujay Man Bajracharya, Yufei Wang, Khush Agrawal, and David Held. Fabricflownet: Bi- manual cloth manipulation with a flow-based policy. In Conference on Robot Learning , pages 192–202. PMLR, 2022

    work page 2022

  6. [6]

    Deep imitation learning for humanoid loco-manipulation through human teleoperation

    Mingyo Seo, Steve Han, Kyutae Sim, Seung Hyeon Bang, Carlos Gonzalez, Luis Sentis, and Yuke Zhu. Deep imitation learning for humanoid loco-manipulation through human teleoperation. In 2023 IEEE-RAS 22nd International Conference on Humanoid Robots (Hu- manoids), pages 1–8. IEEE, 2023

    2023

  7. [7]

    Umi on legs: Making manipulation policies mobile with manipulation-centric whole-body controllers

    Huy Ha, Yihuai Gao, Zipeng Fu, Jie Tan, and Shuran Song. Umi on legs: Making manipulation policies mobile with manipulation-centric whole-body controllers. arXiv preprint arXiv:2407.10353, 2024

    arXiv 2024

  8. [8]

    Humanplus: Humanoid shad- owing and imitation from humans

    Zipeng Fu, Qingqing Zhao, Qi Wu, Gordon Wet- zstein, and Chelsea Finn. Humanplus: Humanoid shad- owing and imitation from humans. arXiv preprint arXiv:2406.10454, 2024

    arXiv 2024

Show all 37 references
  1. [9]

    3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations

    Yanjie Ze, Gu Zhang, Kangning Zhang, Chenyuan Hu, Muhan Wang, and Huazhe Xu. 3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations. In Robotics: Science and Systems (RSS) , 2024

    2024

  2. [10]

    Openvla: An open-source vision-language-action model

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, et al. Openvla: An open-source vision-language-action model. arXiv preprint arXiv:2406.09246, 2024

    2024 arXiv

  3. [11]

    π0: A vision- language-action flow model for general robot control

    Kevin Black, Noah Brown, Danny Driess, Adnan Es- mail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Karol Hausman, Brian Ichter, et al. π0: A vision- language-action flow model for general robot control. arXiv preprint arXiv:2410.24164 , 2024

    arXiv

  4. [12]

    Rdt-1b: a diffusion foundation model for bimanual manipulation

    Songming Liu, Lingxuan Wu, Bangguo Li, Hengkai Tan, Huayu Chen, Zhengyi Wang, Ke Xu, Hang Su, and Jun Zhu. Rdt-1b: a diffusion foundation model for bimanual manipulation. arXiv preprint arXiv:2410.07864 , 2024

    2024 arXiv

  5. [13]

    Carp: Visuomotor policy learning via coarse-to-fine autoregressive prediction

    Zhefei Gong, Pengxiang Ding, Shangke Lyu, Siteng Huang, Mingyang Sun, Wei Zhao, Zhaoxin Fan, and Donglin Wang. Carp: Visuomotor policy learning via coarse-to-fine autoregressive prediction. arXiv preprint arXiv:2412.06782, 2024

    2024 arXiv

  6. [14]

    Fast: Efficient action tokeniza- tion for vision-language-action models

    Karl Pertsch, Kyle Stachowicz, Brian Ichter, Danny Driess, Suraj Nair, Quan Vuong, Oier Mees, Chelsea Finn, and Sergey Levine. Fast: Efficient action tokeniza- tion for vision-language-action models. arXiv preprint arXiv:2501.09747, 2025

    2025 arXiv

  7. [15]

    Ldp: A local diffusion planner for efficient robot navigation and collision avoidance

    Wenhao Yu, Jie Peng, Huanyu Yang, Junrui Zhang, Yifan Duan, Jianmin Ji, and Yanyong Zhang. Ldp: A local diffusion planner for efficient robot navigation and collision avoidance. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 5466–547...

    2024

  8. [16]

    Nomad: Goal masked diffusion policies for navigation and exploration

    Ajay Sridhar, Dhruv Shah, Catherine Glossop, and Sergey Levine. Nomad: Goal masked diffusion policies for navigation and exploration. In 2024 IEEE Interna- tional Conference on Robotics and Automation (ICRA) , pages 63–70. IEEE, 2024

    2024

  9. [17]

    Mo- bile aloha: Learning bimanual mobile manipulation with low-cost whole-body teleoperation

    Zipeng Fu, Tony Z Zhao, and Chelsea Finn. Mo- bile aloha: Learning bimanual mobile manipulation with low-cost whole-body teleoperation. arXiv preprint arXiv:2401.02117, 2024

    2024 arXiv

  10. [18]

    Consistency policy: Accelerated visuo- motor policies via consistency distillation

    Aaditya Prasad, Kevin Lin, Jimmy Wu, Linqi Zhou, and Jeannette Bohg. Consistency policy: Accelerated visuo- motor policies via consistency distillation. Robotics: Science and Systems (RSS) , 2024

    2024

  11. [19]

    Manicm: Real-time 3d diffusion policy via consistency model for robotic manipulation

    Guanxing Lu, Zifeng Gao, Tianxing Chen, Wenxun Dai, Ziwei Wang, and Yansong Tang. Manicm: Real-time 3d diffusion policy via consistency model for robotic manipulation. arXiv preprint arXiv:2406.01586 , 2024

    2024 arXiv

  12. [20]

    Adaflow: Imitation learning with variance-adaptive flow-based policies

    Xixi Hu, Bo Liu, Xingchao Liu, and Qiang Liu. Adaflow: Imitation learning with variance-adaptive flow-based policies. arXiv preprint arXiv:2402.04292 , 2024

    2024 arXiv

  13. [21]

    Bidirectional decoding: Improving action chunking via closed-loop resampling

    Yuejiang Liu, Jubayer Ibn Hamid, Annie Xie, Yoonho Lee, Maximilian Du, and Chelsea Finn. Bidirectional decoding: Improving action chunking via closed-loop resampling. arXiv preprint arXiv:2408.17355 , 2024

    2024 arXiv

  14. [22]

    A mobile manipulation system for one-shot teaching of complex tasks in homes

    Max Bajracharya, James Borders, Dan Helmick, Thomas Kollar, Michael Laskey, John Leichty, Jeremy Ma, Umashankar Nagarajan, Akiyoshi Ochiai, Josh Petersen, et al. A mobile manipulation system for one-shot teaching of complex tasks in homes. In 2020 IEEE International Conference...

    2020

  15. [23]

    Footstep planning for the honda asimo humanoid

    Joel Chestnutt, Manfred Lau, German Cheung, James Kuffner, Jessica Hodgins, and Takeo Kanade. Footstep planning for the honda asimo humanoid. In Proceedings of the 2005 IEEE international conference on robotics and automation , pages 629–634. IEEE, 2005

    2005

  16. [24]

    Optimization based full body control for the atlas robot

    Siyuan Feng, Eric Whitman, X Xinjilefu, and Christo- pher G Atkeson. Optimization based full body control for the atlas robot. In 2014 IEEE-RAS International Conference on Humanoid Robots , pages 120–127. IEEE, 2014

    2014

  17. [25]

    Force strategies for cooperative tasks in multiple mobile manipulation sys- tems

    Oussama Khatib, K Yokoi, K Chang, D Ruspini, R Holm- berg, A Casal, and A Baader. Force strategies for cooperative tasks in multiple mobile manipulation sys- tems. In Robotics Research: The Seventh International Symposium, pages 333–342. Springer, 1996

    1996

  18. [26]

    Towards a personal robotics development platform: Rationale and design of an intrinsically safe personal robot

    Keenan A Wyrobek, Eric H Berger, HF Machiel Van der Loos, and J Kenneth Salisbury. Towards a personal robotics development platform: Rationale and design of an intrinsically safe personal robot. In 2008 IEEE International Conference on Robotics and Automation , pages 2165–2170...

    2008

  19. [27]

    Learning to walk in minutes using massively parallel deep reinforcement learning

    Nikita Rudin, David Hoeller, Philipp Reist, and Marco Hutter. Learning to walk in minutes using massively parallel deep reinforcement learning. In Conference on Robot Learning , pages 91–100. PMLR, 2022

    2022

  20. [28]

    Deep whole-body control: learning a unified policy for manipu- lation and locomotion

    Zipeng Fu, Xuxin Cheng, and Deepak Pathak. Deep whole-body control: learning a unified policy for manipu- lation and locomotion. In Conference on Robot Learning, pages 138–149. PMLR, 2023

    2023

  21. [29]

    Image-based visual servoing of unmanned aerial ma- nipulators for tracking and grasping a moving target

    Yanjie Chen, Yangning Wu, Zhenguo Zhang, Zhiqiang Miao, Hang Zhong, Hui Zhang, and Yaonan Wang. Image-based visual servoing of unmanned aerial ma- nipulators for tracking and grasping a moving target. IEEE Transactions on Industrial Informatics , 19(8): 8889–8899, 2022

    2022

  22. [30]

    Eso-based robust and high-precision tracking control for aerial manipulation

    Huazi Cao, Yongqi Li, Cunjia Liu, and Shiyu Zhao. Eso-based robust and high-precision tracking control for aerial manipulation. IEEE Transactions on Automation Science and Engineering , 21(2):2139–2155, 2023

    2023

  23. [31]

    Millimeter-level pick and peg-in-hole task achieved by aerial manipulator

    Meng Wang, Zeshuai Chen, Kexin Guo, Xiang Yu, Youmin Zhang, Lei Guo, and Wei Wang. Millimeter-level pick and peg-in-hole task achieved by aerial manipulator. IEEE Transactions on Robotics , 2023

    2023

  24. [32]

    Long short-term memory

    S Hochreiter. Long short-term memory. Neural Compu- tation MIT-Press, 1997

    1997

  25. [33]

    Empirical evaluation of gated recurrent neural networks on sequence modeling

    Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555, 2014

    2014 arXiv

  26. [34]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770–778, 2016

    2016

  27. [35]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems , 2017

    2017

  28. [36]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502, 2020

    2010 arXiv

  29. [37]

    Lerobot: State-of-the-art machine learning for real-world robotics in pytorch

    Remi Cadene, Simon Alibert, Alexander Soare, Quentin Gallouedec, Adil Zouitine, and Thomas Wolf. Lerobot: State-of-the-art machine learning for real-world robotics in pytorch. https://github.com/huggingface/lerobot, 2024

    2024

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.