Reinforcement learning-based method for grasping moving object with robotic manipulator
https://doi.org/10.35266/1999-7604-2025-2-8
Abstract
The paper proposes a method for controlling a robotic manipulator using reinforcement learning with deep neural networks to grasp moving objects on a conveyor belt. Unlike tasks involving static objects, this problem requires the consideration of dynamic factors, which significantly complicates the control process. The paper provides a detailed description of the physical and kinematic modeling of the manipulator, as well as the integration of manipulator and object parameters into the neural network structure. The method is tested in the PyBullet physics simulation environment.
References
1. Shuzhi S. G., Hang C. C., Woon L. C. Adaptive neural network control of robot manipulators in task space // IEEE transactions on industrial electronics. 1997. No. 6. P. 746–752.
2. Pane Y. P., Nageshrao S. P., Kober J. et al. Reinforcement learning based compensation methods for robot manipulators // Engineering Applications of Artificial Intelligence. 2019. Vol. 78. P. 236–247.
3. Sekkat H., Tigani S., Saadane R. et al. Vision-based robotic arm control algorithm using deep reinforcement learning for autonomous objects grasping // Applied Sciences. 2021. No. 11. P. 7917.
4. Marturi N., Kopicki M., Rastegarpanah A. et al. Dynamic grasp and trajectory planning for moving objects // Autonomous Robots. 2019. No. 43. P. 1241–1256.
5. Reed A., Albin D., Pasricha A. et al. Transformer-based Learning Models of Dynamical Systems for Robotic State Prediction. 2024. https://doi.org/10.21203/rs.3.rs-3919154/v1.
6. Цао И. Метод визуально управляемого захвата 7-степенного манипулятора на основе обучения с подкреплением // Вестник кибернетики. 2025. Т. 24, № 1. С. 31–38.
7. Lillicrap T. P., Hunt J. J., Pritzel A. et al. Continuous control with deep reinforcement learning // arXiv preprint arXiv:1509.02971. 2015.
8. Fujimoto S., Hoof H., Meger D. Addressing function approximation error in actor-critic methods // Proceedings of the 35th International conference on machine learning, 2018, Stockholm. PMLR, 2018. P. 1587–1596.
9. Haarnoja T., Zhou A., Abbeel P. et al. Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor // Proceedings of the 35th International conference on machine learning, 2018, Stockholm. PMLR, 2018. P. 1861–1870.
10. Schulman J., Wolski F., Dhariwal P. et al. Proximal policy optimization algorithms // arXiv preprint arXiv: 1707.06347. 2017.
Review
For citations:
Cao Y. Reinforcement learning-based method for grasping moving object with robotic manipulator. Proceedings in Cybernetics. 2025;24(2):66-73. (In Russ.) https://doi.org/10.35266/1999-7604-2025-2-8
Views:
46
Email this article 