arxiv: 2604.22235 · v1 · submitted 2026-04-24 · 💻 cs.RO · cs.AI· cs.LG

Recognition: unknown

Learning-augmented robotic automation for real-world manufacturing

Yunho Kim , Quan Nguyen , Taewhan Kim , Youngjin Heo , Joonho Lee

Authors on Pith no claims yet

Pith reviewed 2026-05-08 11:20 UTC · model grok-4.3

classification 💻 cs.RO cs.AIcs.LG

keywords learning-based controlindustrial roboticssafety monitoringmanufacturing automationdeformable manipulationreal-world deploymentcable insertionsoldering

0 comments

The pith

A hybrid system of learned controllers and a neural 3D safety monitor lets industrial robots automate deformable cable insertion and soldering on a live line after under 20 minutes of real data per task.

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

The paper establishes that learning-based methods can move from lab settings into sustained factory operation by combining task-specific controllers with a neural safety layer that watches the 3D workspace. It demonstrates this on an electric-motor production line where the robots insert deformable cables and perform soldering, tasks previously done by hand. With less than twenty minutes of on-site data collection the system ran uninterrupted for five hours and ten minutes, completed 108 motors, required no physical fencing, and passed 99.4 percent of final quality checks while matching human speed and cutting variability. A reader would care because the result directly addresses the brittleness of fixed robot programs to small environmental shifts and shows a concrete route to safer, more flexible automation around people.

Core claim

The authors claim that a hybrid Learning-Augmented Robotic Automation system, which fuses learned task controllers with a neural 3D safety monitor, can be inserted into existing industrial robot workflows to handle deformable cable insertion and soldering. On an electric-motor line the setup used under 20 minutes of real-world data per task, operated continuously for 5 hours 10 minutes, produced 108 motors without physical fencing, reached a 99.4 percent pass rate on product-level quality-control tests, maintained near-human takt time, and reduced variability in solder-joint quality and cycle time.

What carries the argument

The hybrid integration of learned task controllers for manipulation with a neural 3D safety monitor that provides real-time hazard awareness without physical barriers.

If this is right

Fixed waypoint programming can be augmented for tasks that involve flexible or variable objects such as cables.
Industrial robots can operate without enclosing safety fences while remaining safe around human workers on active lines.
High consistency in manufacturing quality and timing is achievable after only short periods of on-site data collection.
Variability in both product quality and process timing can be lowered relative to fully manual execution of the same tasks.

Where Pith is reading between the lines

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

If the safety monitor generalizes across product variants, factories could shorten changeover times by retraining controllers in minutes rather than hours.
The approach may extend to other sectors with deformable materials or high-mix production where fixed programs currently fail.
Over longer horizons the system would likely need mechanisms for gradual online adaptation to tool wear or fixture drift.
Similar hybrid safety layers could be tested in logistics or assembly tasks where robots and people share space.

Load-bearing premise

The learned controllers trained on limited site-specific data together with the neural safety monitor will continue to perform reliably and safely under ongoing production variability, environmental changes, and durations longer than the single five-hour test.

What would settle it

A multi-shift or multi-day run on the same line that records either a quality pass rate falling below 99 percent, frequent safety-monitor interventions, or increasing cycle-time variability would falsify the claim of sustained reliable operation.

Figures

Figures reproduced from arXiv: 2604.22235 by Joonho Lee, Quan Nguyen, Taewhan Kim, Youngjin Heo, Yunho Kim.

**Figure 1.** Figure 1: Production line deployment. (a) Overview of the workstation deployed on the factory floor. (b) Production process and division of labor. A human worker loads motor cores and consumables into the cell (1). The robot then executes the cable insertion-and-soldering sequence, including motor placement/orientation, sequential insertion of the three-phase power cables, soldering, and tip cleaning (2–5). Finally,… view at source ↗

**Figure 2.** Figure 2: Learning-augmented robotic automation. The software stack comprises modular task controllers and a safety monitoring module. Learned components are integrated at the task and safety levels to overcome the limitation of conventional automation. Results System Overview We structured the automation pipeline as a sequence of modular manipulation tasks coordinated in a cyclic workflow (Figure 1b). Learning-Augm… view at source ↗

**Figure 3.** Figure 3: Behavior of learned controllers. (a) Visual servoing controller uses backgroundremoved motor-core images as input (i) and iteratively adjusts the end-effector pose at each inference step until the target view is reached (ii). (b) Imitation learning controller takes stereo RGB inputs from the worktable cameras. A lightweight mask predictor extracts the three PCB holes, and the policy is conditioned on the … view at source ↗

**Figure 4.** Figure 4: Production line deployment results. (a) Production outcomes: (a-i) single-station cycle time (pick, insert, solder, tip clean) for a human worker vs. the robot; (a-ii) total output during the operation; (a-iii) downstream product-level QC pass rate. (b) Cumulative throughput projection for an 8-hour shift. “Robot alone” is computed from the nominal cycle time, whereas “Robot between humans” uses the effect… view at source ↗

**Figure 5.** Figure 5: Comparison with task controller variants. (a) view at source ↗

**Figure 6.** Figure 6: 3D Safety Monitoring with Speed Modulation. (a) view at source ↗

read the original abstract

Industrial robots are widely used in manufacturing, yet most manipulation still depends on fixed waypoint scripts that are brittle to environmental changes. Learning-based control offers a more adaptive alternative, but it remains unclear whether such methods, still mostly confined to laboratory demonstrations, can sustain hours of reliable operation, deliver consistent quality, and behave safely around people on a live production line. Here we present Learning-Augmented Robotic Automation, a hybrid system that integrates learned task controllers and a neural 3D safety monitor into conventional industrial workflows. We deployed the system on an electric-motor production line to automate deformable cable insertion and soldering under real manufacturing constraints, a step previously performed manually by human workers. With less than 20 min of real-world data per task, the system operated continuously for 5 h 10 min, producing 108 motors without physical fencing and achieving a 99.4% pass rate on product-level quality-control tests. It maintained near-human takt time while reducing variability in solder-joint quality and cycle time. These results establish a practical pathway for extending industrial automation with learning-based methods.

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.

Desk Editor's Note private letter to a colleague

The paper delivers a concrete 5-hour unfenced deployment on a live motor line with 99.4% pass rate after minimal data, but one run leaves generalization untested.

read the letter

Colleague, the main thing here is the real deployment. They ran a hybrid system of learned controllers plus a neural safety monitor on an actual electric-motor production line for 5 hours 10 minutes, handling deformable cable insertion and soldering on 108 motors without fencing. It hit 99.4% on product QC tests, kept cycle times near human levels, and cut variability in solder quality. That sustained operation with under 20 minutes of site-specific data per task is the new piece compared to typical lab demos. The hybrid setup makes sense for fitting into existing industrial robots while adding adaptability and safety around people. Reporting those specific metrics from live production gives it more weight than pure simulation results. The soft spots are straightforward. Everything rests on a single continuous trial. No replication across days, no tests with temperature or lighting drift, no cable batch variation, and no numbers on how often the safety monitor stepped in or any near-miss events. The abstract gives no baselines against fixed scripts or statistical breakdown of the outcomes, so the 99.4% figure is hard to interpret without more context. The assumption that the learned parts will hold up under ongoing manufacturing variability is the weakest link, and the single run does not address it. This is for robotics people and manufacturing engineers who want to see learning methods actually running on a line. It shows a possible path forward but will need more runs and failure analysis to convince skeptics. I would send it to peer review because the empirical result is worth referee scrutiny even if the robustness claims need strengthening.

Referee Report

1 major / 2 minor

Summary. The paper presents Learning-Augmented Robotic Automation, a hybrid system that combines learned task controllers with a neural 3D safety monitor integrated into conventional industrial robot workflows. It reports a real-world deployment automating deformable cable insertion and soldering on an electric-motor production line, a task previously done manually. With under 20 minutes of site-specific real-world data per task, the system ran continuously for 5 hours 10 minutes without physical fencing, produced 108 motors, achieved a 99.4% pass rate on product-level quality-control tests, maintained near-human takt time, and reduced variability in solder-joint quality and cycle time. These results are positioned as establishing a practical pathway for learning-based methods in manufacturing.

Significance. If the reported performance is shown to generalize, the work would be significant as one of the few documented cases of a learning-augmented controller plus neural safety system sustaining multi-hour operation on a live, unfenced production line with deformable objects. The quantitative outcomes from physical deployment (motor count, pass rate, takt time) provide concrete evidence that could help bridge laboratory learning methods to factory constraints, particularly the safety and reliability requirements that have limited adoption.

major comments (1)

[Deployment results (as summarized in abstract and experimental section)] The central claim that the hybrid system provides a 'practical pathway' for real-world manufacturing rests on results from a single continuous 5 h 10 min deployment trial. No replication across days or shifts, no induced or observed environmental drift (lighting, temperature, cable batch variation), and no statistics on safety-monitor interventions, near-miss events, or failure modes are reported. This directly limits evaluation of robustness under ongoing production variability, which is load-bearing for the broader assertion.

minor comments (2)

[Abstract and results] The abstract states 'near-human takt time' and 'reduced variability' without providing the numerical baseline human takt time, the exact definition of cycle-time variability, or the statistical test used; these should be stated explicitly with values in the results section.
[Methods] Details on the data-collection protocol (how the <20 min of real-world data were gathered, including any human demonstrations or teleoperation), the precise architecture of the neural 3D safety monitor, and the criteria for the 99.4% QC pass rate are referenced but not fully elaborated; adding these would strengthen reproducibility.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive assessment and the opportunity to address the scope of our deployment claims. We respond to the major comment below.

read point-by-point responses

Referee: [Deployment results (as summarized in abstract and experimental section)] The central claim that the hybrid system provides a 'practical pathway' for real-world manufacturing rests on results from a single continuous 5 h 10 min deployment trial. No replication across days or shifts, no induced or observed environmental drift (lighting, temperature, cable batch variation), and no statistics on safety-monitor interventions, near-miss events, or failure modes are reported. This directly limits evaluation of robustness under ongoing production variability, which is load-bearing for the broader assertion.

Authors: We agree that the reported results derive from a single continuous 5 h 10 min trial on a live production line and that this constrains statistical claims about robustness to ongoing variability. The trial incorporated whatever natural environmental fluctuations occurred during that window, but we neither induced nor systematically measured drifts in lighting, temperature, or cable batches. No safety-monitor interventions or near-miss events requiring human response were observed, which is why granular statistics on those quantities were not included. We will revise the manuscript by adding an explicit Limitations subsection that states the single-trial nature of the evidence, describes the conditions present during the run, and qualifies the 'practical pathway' language as an initial demonstration of sustained operation rather than a comprehensive robustness validation. This change will be reflected in the abstract and experimental discussion as well. revision: partial

Circularity Check

0 steps flagged

No circularity; empirical deployment results are direct measurements

full rationale

The paper presents a hybrid learned-controller plus neural safety system deployed on a live motor production line. Its central claims rest on measured outcomes from a single 5-hour physical run (108 motors, 99.4% QC pass rate, <20 min site-specific data) rather than any mathematical derivation, first-principles prediction, or fitted-parameter renaming. No equations, uniqueness theorems, or self-citation chains are invoked to derive the reported performance; the numbers are obtained by direct observation of the physical system. This matches the default expectation for an empirical robotics paper and yields no circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the empirical success of integrating learned controllers with a safety monitor under real constraints rather than on new theoretical constructs; no free parameters or invented entities are introduced in the abstract.

axioms (1)

domain assumption Learned controllers trained on limited site-specific data can reliably handle deformable objects and variable conditions in manufacturing
Invoked to justify the <20 min data collection and sustained operation claim

pith-pipeline@v0.9.0 · 5497 in / 1286 out tokens · 31081 ms · 2026-05-08T11:20:17.564039+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

54 extracted references · 7 canonical work pages · 4 internal anchors

[1]

Vysocky, P

A. Vysocky, P. Novak, Human-robot collaboration in industry,MM Science Journal903– 906 (2016)

2016
[2]

El Zaatari, M

S. El Zaatari, M. Marei, W. Li, Z. Usman, Cobot programming for collaborative industrial tasks: An overview,Robotics and Autonomous Systems162–180 (2019)

2019
[3]

Graetz, G

G. Graetz, G. Michaels, Robots at work,Review of economics and statistics753–768 (2018)

2018
[4]

Accessed: June 1, 2025

Universal Robots, Universal robots user manual, https://www.universal- robots.com/manuals/EN/HTML/MainLanding/Content/Landingpages/mainlanding.htm (2024). Accessed: June 1, 2025

2024
[5]

Accessed: June 1, 2025

Neuromeka, Conty documentation, http://docs.neuromeka.com/3.2.0/en/Conty/conty/ (2024). Accessed: June 1, 2025

2024
[6]

Y . Cong, R. Chen, B. Ma, H. Liu, D. Hou, C. Yang, A comprehensive study of 3-d vision- based robot manipulation,IEEE Transactions on Cybernetics1682–1698 (2021)

2021
[7]

T. Z. Zhao, V . Kumar, S. Levine, C. Finn, Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware,Proceedings of Robotics: Science and Systems(Daegu, Republic of Korea, 2023)

2023
[8]

Z. Fu, T. Z. Zhao, C. Finn, Mobile aloha: Learning bimanual mobile manipulation with low-cost whole-body teleoperation,Conference on Robot Learning (CoRL)(2024)

2024
[9]

C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burchfiel, R. Tedrake, S. Song, Diffu- sion policy: Visuomotor policy learning via action diffusion,The International Journal of Robotics Research(2024). 37

2024
[10]

J. W. Kim, J.-T. Chen, P. Hansen, L. X. Shi, A. Goldenberg, S. Schmidgall, P. M. Scheikl, A. Deguet, B. M. White, D. R. Tsai, others, Srt-h: A hierarchical framework for autonomous surgery via language-conditioned imitation learning,Science roboticsp. eadt5254 (2025)

2025
[11]

J. Luo, Z. Hu, C. Xu, Y . L. Tan, J. Berg, A. Sharma, S. Schaal, C. Finn, A. Gupta, S. Levine, Serl: A software suite for sample-efficient robotic reinforcement learning,2024 IEEE In- ternational Conference on Robotics and Automation (ICRA), 16961–16969 (IEEE, 2024)

2024
[12]

J. Luo, C. Xu, J. Wu, S. Levine, Precise and dexterous robotic manipulation via human-in- the-loop reinforcement learning,Science Roboticsp. eads5033 (2025)

2025
[13]

K. Lei, H. Li, D. Yu, Z. Wei, L. Guo, Z. Jiang, Z. Wang, S. Liang, H. Xu, Rl-100: Performant robotic manipulation with real-world reinforcement learning,arXiv preprint arXiv:2510.14830(2025)

work page arXiv 2025
[14]

Zitkovich, T

B. Zitkovich, T. Yu, S. Xu, P. Xu, T. Xiao, F. Xia, J. Wu, P. Wohlhart, S. Welker, A. Wahid, Q. Vuong, V . Vanhoucke, H. Tran, R. Soricut, A. Singh, J. Singh, P. Sermanet, P. R. Sanketi, G. Salazar, M. S. Ryoo, K. Reymann, K. Rao, K. Pertsch, I. Mordatch, H. Michalewski, Y . Lu, S. Levine, L. Lee, T.-W. E. Lee, I. Leal, Y . Kuang, D. Kalashnikov, R. Julia...

2023
[15]

M. J. Kim, K. Pertsch, S. Karamcheti, T. Xiao, A. Balakrishna, S. Nair, R. Rafailov, E. P. Foster, P. R. Sanketi, Q. Vuong, T. Kollar, B. Burchfiel, R. Tedrake, D. Sadigh, S. Levine, P. Liang, C. Finn, OpenVLA: An open-source vision-language-action model,8th Annual Conference on Robot Learning(2024)

2024
[16]

Black, N

K. Black, N. Brown, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fusai, L. Groom, K. Haus- man, B. Ichter, S. Jakubczak, T. Jones, L. Ke, S. Levine, A. Li-Bell, M. Mothukuri, S. Nair, K. Pertsch, L. X. Shi, L. Smith, J. Tanner, Q. Vuong, A. Walling, H. Wang, U. Zhilinsky, π0: A vision-language-action flow model for general robot control,Robotics: Science an...

2025
[17]

Black, N

K. Black, N. Brown, J. Darpinian, K. Dhabalia, D. Driess, A. Esmail, M. R. Equi, C. Finn, N. Fusai, M. Y . Galliker, D. Ghosh, L. Groom, K. Hausman, b. ichter, S. Jakubczak, T. Jones, L. Ke, D. LeBlanc, S. Levine, A. Li-Bell, M. Mothukuri, S. Nair, K. Pertsch, A. Z. Ren, L. X. Shi, L. Smith, J. T. Springenberg, K. Stachowicz, J. Tanner, Q. Vuong, H. Walke...

2025
[18]

GR00T N1: An Open Foundation Model for Generalist Humanoid Robots

J. Bjorck, F. Casta ˜neda, N. Cherniadev, X. Da, R. Ding, L. Fan, Y . Fang, D. Fox, F. Hu, S. Huang, others, Gr00t n1: An open foundation model for generalist humanoid robots, arXiv preprint arXiv:2503.14734(2025)

work page internal anchor Pith review arXiv 2025
[19]

G. R. Team, S. Abeyruwan, J. Ainslie, J.-B. Alayrac, M. G. Arenas, T. Armstrong, A. Bal- akrishna, R. Baruch, M. Bauza, M. Blokzijl, others, Gemini robotics: Bringing ai into the physical world,arXiv preprint arXiv:2503.20020(2025). 39

work page internal anchor Pith review arXiv 2025
[20]

H. Li, K. Lei, S. Zang, K. Hu, Y . Liang, B. An, X. Li, H. Xu, Failure-aware rl: Reli- able offline-to-online reinforcement learning with self-recovery for real-world manipula- tion,arXiv preprint arXiv:2601.07821(2026)

work page arXiv 2026
[21]

Y . Ma, Z. Song, Y . Zhuang, J. Hao, I. King, A survey on vision-language-action models for embodied ai,arXiv preprint arXiv:2405.14093(2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024
[22]

Tsuji, Y

T. Tsuji, Y . Kato, G. Solak, H. Zhang, T. Petriˇc, F. Nori, A. Ajoudani, A survey on imitation learning for contact-rich tasks in robotics,The International Journal of Robotics Research p. 02783649261417694 (2025)

2025
[23]

Dreczkowski, P

K. Dreczkowski, P. Vitiello, V . V osylius, E. Johns, Learning a thousand tasks in a day, Science Roboticsp. eadv7594 (2025)

2025
[24]

N. R. Arachchige, Z. Chen, W. Jung, W. C. Shin, R. Bansal, P. Barroso, Y . H. He, Y . C. Lin, B. Joffe, S. Kousik, D. Xu, Sail: Faster-than-demonstration execution of imitation learning policies,9th Annual Conference on Robot Learning(2025)

2025
[25]

J. Lee, Y . Kim, S. Kim, Q. Nguyen, Y . Heo, Learning fast, tool-aware collision avoidance for collaborative robots,IEEE Robotics and Automation Letters(2025)

2025
[26]

S. Peng, M. Niemeyer, L. Mescheder, M. Pollefeys, A. Geiger, Convolutional occupancy networks,European Conference on Computer Vision, 523–540 (Springer, 2020)

2020
[27]

International Organization for Standardization, Robots and robotic devices — collaborative robots,Technical Specification ISO/TS 15066:2016, ISO, Geneva, Switzerland (2016)

2016
[28]

J. A. Marvel, R. Norcross, Implementing speed and separation monitoring in collaborative robot workcells,Robotics and computer-integrated manufacturing144–155 (2017). 40

2017
[29]

Accessed: February 11, 2026

Universal Robots, Universal robots safety sensors, https://www.universal- robots.com/marketplace/products/01tP40000071NhmIAE/ (2026). Accessed: February 11, 2026

2026
[30]

Karagiannis, N

P. Karagiannis, N. Kousi, G. Michalos, K. Dimoulas, K. Mparis, D. Dimosthenopoulos, ¨O. Tokc ¸alar, T. Guasch, G. P. Gerio, S. Makris, Adaptive speed and separation monitoring based on switching of safety zones for effective human robot collaboration,Robotics and Computer-Integrated Manufacturingp. 102361 (2022)

2022
[31]

E. Johns, Coarse-to-fine imitation learning: Robot manipulation from a single demonstra- tion,2021 IEEE international conference on robotics and automation (ICRA), 4613–4619 (IEEE, 2021)

2021
[32]

C. Yu, Z. Cai, H. Pham, Q.-C. Pham, Siamese convolutional neural network for sub- millimeter-accurate camera pose estimation and visual servoing,2019 IEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems (IROS), 935–941 (IEEE, 2019)

2019
[33]

N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R ¨adle, C. Rolland, L. Gustafson, E. Mintun, J. Pan, K. V . Alwala, N. Carion, C.-Y . Wu, R. Girshick, P. Dol- lar, C. Feichtenhofer, SAM 2: Segment anything in images and videos,The Thirteenth International Conference on Learning Representations(2025)

2025
[34]

S. Liu, Z. Zeng, T. Ren, F. Li, H. Zhang, J. Yang, Q. Jiang, C. Li, J. Yang, H. Su, others, Grounding dino: Marrying dino with grounded pre-training for open-set object detection, European conference on computer vision, 38–55 (Springer, 2024)

2024
[35]

Oquab, T

M. Oquab, T. Darcet, T. Moutakanni, H. V . V o, M. Szafraniec, V . Khalidov, P. Fernandez, D. HAZIZA, F. Massa, A. El-Nouby, M. Assran, N. Ballas, W. Galuba, R. Howes, P.-Y . Huang, S.-W. Li, I. Misra, M. Rabbat, V . Sharma, G. Synnaeve, H. Xu, H. Jegou, J. Mairal, 41 P. Labatut, A. Joulin, P. Bojanowski, DINOv2: Learning robust visual features without su...

2024
[36]

C. Chi, Z. Xu, C. Pan, E. Cousineau, B. Burchfiel, S. Feng, R. Tedrake, S. Song, Universal manipulation interface: In-the-wild robot teaching without in-the-wild robots,Proceedings of Robotics: Science and Systems (RSS)(2024)

2024
[37]

Ronneberger, P

O. Ronneberger, P. Fischer, T. Brox, U-net: Convolutional networks for biomedical im- age segmentation,International Conference on Medical image computing and computer- assisted intervention, 234–241 (Springer, 2015)

2015
[38]

H. Kim, Y . Ohmura, Y . Kuniyoshi, Gaze-based dual resolution deep imitation learning for high-precision dexterous robot manipulation,IEEE Robotics and Automation Letters 1630–1637 (2021)

2021
[39]

Vaswani, N

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, I. Polo- sukhin, Attention is all you need,Advances in neural information processing systems30 (2017)

2017
[40]

D. Ko, W. K. Chung, A backup control barrier function approach for safety-critical control of mechanical systems under multiple constraints,IEEE/ASME Transactions on Mecha- tronics4460–4471 (2025)

2025
[41]

D. Lee, D. Ko, W. K. Chung, K. Kim, Quadratic programming-based task scaling for safe and passive robot arm teleoperation,IEEE/ASME Transactions on Mechatronics1937– 1945 (2022)

1945
[42]

Y . J. Heo, D. Kim, W. Lee, H. Kim, J. Park, W. K. Chung, Collision detection for industrial collaborative robots: A deep learning approach,IEEE Robotics and Automation Letters 740–746 (2019). 42

2019
[43]

J. Yang, J. J. Liu, Y . Li, Y . Khaky, D. Pathak, Deep reactive policy: Learning reactive manipulator motion planning for dynamic environments,9th Annual Conference on Robot Learning(2025)

2025
[44]

Oleynikova, Z

H. Oleynikova, Z. Taylor, M. Fehr, R. Siegwart, J. Nieto, V oxblox: Incremental 3d eu- clidean signed distance fields for on-board mav planning,2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 1366–1373 (IEEE, 2017)

2017
[45]

Sundaralingam, S

B. Sundaralingam, S. K. S. Hari, A. Fishman, C. Garrett, K. Van Wyk, V . Blukis, A. Mil- lane, H. Oleynikova, A. Handa, F. Ramos, others, Curobo: Parallelized collision-free robot motion generation,2023 IEEE International Conference on Robotics and Automa- tion (ICRA), 8112–8119 (IEEE, 2023)

2023
[46]

L. Zhu, M. Menon, M. Santillo, G. Linkowski, Occlusion handling for industrial robots, 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 10663–10668 (IEEE, 2020)

2020
[47]

Macklin, Warp: A high-performance python framework for gpu simulation and graph- ics,https://github.com/nvidia/warp(2022)

M. Macklin, Warp: A high-performance python framework for gpu simulation and graph- ics,https://github.com/nvidia/warp(2022). NVIDIA GPU Technology Con- ference (GTC)

2022
[48]

M. T. Villasevil, A. Simeonov, Z. Li, A. Chan, T. Chen, A. Gupta, P. Agrawal, Recon- ciling reality through simulation: A real-to-sim-to-real approach for robust manipulation, Robotics: Science and Systems(2024)

2024
[49]

Isaac Lab: A GPU-Accelerated Simulation Framework for Multi-Modal Robot Learning

M. Mittal, P. Roth, J. Tigue, A. Richard, O. Zhang, P. Du, A. Serrano-Mu ˜noz, X. Yao, R. Zurbr¨ugg, N. Rudin, L. Wawrzyniak, M. Rakhsha, A. Denzler, E. Heiden, A. Borovicka, O. Ahmed, I. Akinola, A. Anwar, M. T. Carlson, J. Y . Feng, A. Garg, R. Gasoto, L. Gulich, Y . Guo, M. Gussert, A. Hansen, M. Kulkarni, C. Li, W. Liu, V . Makoviychuk, G. Malczyk, 43...

work page internal anchor Pith review arXiv 2025
[50]

S. Ross, G. Gordon, D. Bagnell, A reduction of imitation learning and structured prediction to no-regret online learning,Proceedings of the fourteenth international conference on ar- tificial intelligence and statistics, 627–635 (JMLR Workshop and Conference Proceedings, 2011)

2011
[51]

Kelly, C

M. Kelly, C. Sidrane, K. Driggs-Campbell, M. J. Kochenderfer, Hg-dagger: Interactive imitation learning with human experts,2019 International Conference on Robotics and Automation (ICRA), 8077–8083 (IEEE, 2019)

2019
[52]

Thumm, J

J. Thumm, J. Balletshofer, L. Maglanoc, L. Muschal, M. Althoff, A general safety framework for autonomous manipulation in human environments,arXiv preprint arXiv:2412.10180(2024). 44

work page arXiv 2024
[53]

Haddadin, S

S. Haddadin, S. Haddadin, A. Khoury, T. Rokahr, S. Parusel, R. Burgkart, A. Bicchi, A. Albu-Sch¨affer, On making robots understand safety: Embedding injury knowledge into control,The International Journal of Robotics Research1578–1602 (2012)

2012
[54]

wrist image

Y . Zhou, C. Barnes, J. Lu, J. Yang, H. Li, On the continuity of rotation representations in neural networks,Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 5745–5753 (2019). 45 Supplementary Materials Section S1. Implementation Details for Zero-shot Mask Tracker Section S2. Implementation Details for VLA baseline Sectio...

2019