Reinforcement Learning for Adaptive Grasping of Objects by a Collaborative Robot: from Simulation to a Real System

Polina  Lazareva,Nikolay  Gurko

ПРЕПРИНТ

Эта статья является препринтом и не была отрецензирована.
О результатах, изложенных в препринтах, не следует сообщать в СМИ как о проверенной информации.

Reinforcement Learning for Adaptive Grasping of Objects by a Collaborative Robot: from Simulation to a Real System

P. Lazareva,

N. Gurko

2026-01-20

https://doi.org/10.24108/preprints-3114305

This study presents a method for training a collaborative robot to perform adaptive object grasping using reinforcement learning and a curriculum learning strategy, followed by transferring the resulting high-level policy from simulation to a real robotic system. A digital model of the Rozum Pulse 75 robot was created in Unity with ML-Agents, where the grasping task was decomposed into a sequence of progressively more complex subtasks involving coarse positioning, orientation alignment, and grasp execution. To bridge the gap between simulation and reality, the high-level policy operates on target poses of the end-effector, while low-level dynamics are handled by the robot’s built-in controller. The trained policy achieved an 87% success rate in simulation and an 82% success rate on the physical robot, without additional fine-tuning. The results confirm that the proposed training pipeline provides stable learning in simulation and enables effective Sim2Real transfer for robotic manipulation tasks.

Ссылка для цитирования:

Lazareva P., Gurko N. 2026. Reinforcement Learning for Adaptive Grasping of Objects by a Collaborative Robot: from Simulation to a Real System. PREPRINTS.RU. https://doi.org/10.24108/preprints-3114305

Список литературы

1. K. Chatzilygeroudis, V. Vassiliades, F. Stulp, S. Calinon, and J.-B. Mouret, IEEE Trans. Robot. 36, 328–347 (2020).

2. Y. Chebotar et al., “Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experience,” in 2019 International Conference on Robotics and Automation (ICRA) (IEEE, Montreal, 2019), pp. 8973–8979.

3. A. Handa et al., “DeXtreme: Transfer of Agile In-hand Manipulation from Simulation to Reality,” in 2023 IEEE International Conference on Robotics and Automation (ICRA) (IEEE, London, 2023), pp. 5977–5984.

4. A. Allshire et al., “Transferring Dexterous Manipulation from GPU Simulation to a Remote Real-World TriFinger,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE, Kyoto, 2022), pp. 11802–11809.

5. E. Salvato, G. Fenu, E. Medvet, and F. A. Pellegrino, IEEE Access 9, 153171–153187 (2021).

6. J. Josifovski, M. Malmir, N. Klarmann, B. L. Žagar, N. Navarro-Guerrero, and A. Knoll, “Analysis of Randomization Effects on Sim2Real Transfer in Reinforcement Learning for Robotic Manipulation Tasks,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE, Kyoto, 2022), pp. 10193–10200.

7. J. Huber, F. Hélénon, H. Watrelot, F. B. Amar, and S. Doncieux, “Domain Randomization for Sim2real Transfer of Automatically Generated Grasping Datasets,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) (IEEE, Yokohama, 2024), pp. 4112–4118.

8. W. Zhao, J. P. Queralta, and T. Westerlund, “Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: A Survey,” in 2020 IEEE Symposium Series on Computational Intelligence (SSCI) (IEEE, Canberra, 2020), pp. 737–744.

9. C. Wang et al., “DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (IEEE, Long Beach, 2019), pp. 3338–3347.

10. Y. He, W. Sun, H. Huang, J. Liu, H. Fan, and J. Sun, “PVN3D: A Deep Point-Wise 3D Keypoints Voting Network for 6DoF Pose Estimation,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (IEEE, Seattle, 2020), pp. 11629–11638.

11. M. Yan, I. Frosio, S. Tyree, and J. Kautz, “Sim-to-Real Transfer of Accurate Grasping with Eye-In-Hand Observations and Continuous Control,” arXiv:1712.03303 (2017).

12. J. L. Elman, “Learning and development in neural networks: The importance of starting small,” Cognition 48, 71–99 (1993).

13. C. Florensa, D. Held, M. Wulfmeier, M. Zhang, and P. Abbeel, “Reverse Curriculum Generation for Reinforcement Learning,” in Proceedings of the 1st Annual Conference on Robot Learning (PMLR, 2017), pp. 482–495.

14. T. Matiisen, A. Oliver, T. Cohen, and J. Schulman, “Teacher-Student Curriculum Learning,” IEEE Trans. Neural Netw. Learn. Syst. 31, 3732–3740 (2020).

ПРЕПРИНТ

ИНФОРМАЦИЯ

Использование куки-файлов