INCREASED EFFECTIVE OF CONTROL ROBOT  USING REINFORCEMENT COMBINE WITH DEEP LEARNING

Lương Thị Thảo Hiếu; Phạm Thị Thùy; Nguyễn Khắc Hiệp

doi:10.34238/tnu-jst.7733

INCREASED EFFECTIVE OF CONTROL ROBOT USING REINFORCEMENT COMBINE WITH DEEP LEARNING

About this article

Received: 13/04/23 Revised: 24/05/23 Published: 24/05/23

Authors

1. Luong Thi Thao Hieu , University of Economic and Technical Industries
2. Pham Thi Thuy, University of Economic and Technical Industries
3. Nguyen Khac Hiep, University of Economic and Technical Industries

Abstract

Although deep learning can solve problems that cannot be done by tradition machine learning algorithms, it requires huge amount of data, which is not always available in control problems. Reinforcement learning is a good solution in robot control. In reinforcement learning, the data is generated when the agent interacts with environment. Along with the develop of noron networks, many researcher have focused on combine noron network with reinforcement learning to create deep reinforcement learning. In this paper, we propose a new deep reinforcement learning model based on the improvement of the traditional Deep Q Learning algorithm by combining techniques: Fixed_Q Target, Double Deep Q, Prioritized Experience Replay with CNN network (VGG16), apply to control the robot with state and agents designed by using Unity ML-Agents. The proposed model was tested, compared to the original model on the simulation environment. The results show that the proposed method fix the overestimation q value and fast convergence.

Keywords

Reinforcement learning; Deep reinforcement learning; Robotic manipulation control; DQN; VGG16

Full Text:

PDF (Tiếng Việt)

References

[1] J. Kober, J. A. Bagnell, and J. Peters, “Reinforcement Learning in Robotics: A Survey,” International Journal of Robotics Research, vol. 32, no. 11, pp. 1238-1274, 2013.

[2] R. Liu, F. Nageotte, and P. Zanne, “Deep reinforcement learning for control of robotic manipulation: A review,” Journal of Robotics, vol. 10, no. 1, pp. 1-14, 2020.

[3] A. Barzegar and D.-J. Lee, “Deep Reinforcement Learning-Based Adaptive Controller for Trajectory Tracking and Altitude Control of an Aerial Robot,” Applied Sciences, vol. 9, pp. 1-8, 2022.

[4] V. Minh, K. Kavukcuoglu, D. Silver, and A. Graves, “Playing Atari with Deep Reinforcement Learning,” Noron Information Proccessing Systems, vol. 26, pp. 2674-2682, 2013.

[5] V. Mnih, K. Kavukcuoglu, D. Silver, and A. Graves, “Human level Control through Deep Reinforcement Learning,” Nature, vol. 518, pp. 529-533, 2015.

[6] H. V. Hasselt, A. Guez, and D. Silver, “Deep Reinforcement Learning with Double Q-Learning,” in AAAI Conference on Artifical Intelligence (AAAI), 2016, pp. 2094-2100.

[7] C. A. W. Y. Wang, “A Deep Reinforcement Learning Method for Robot Dynamic Path Planning,” IEEE Access, vol. 8, pp. 33817-33826, 2020.

[8] Ó. Pérez-Gil, R. Barea, E. López-Guillén, L. M. Bergasa, C. Gómez-Huélamo, R. Gutiérrez, and A. Díaz-Díaz, “Deep reinforcement learning based control for Autonomous Vehicles in CARLA,” Multimed Tools Appl, vol. 81, no. 3, pp. 3553–3576, 2022

[9] T. Schaul, “Prioritzed Experience Replay,” in ICLR, 2016, pp.1-8.

[10] K. Simonyan and A. Zisserman, “Very deep convolutional network for large-scale image recognition,” in The 3rd International Conference on Learning Representations(ICLR2015), 2015, pp. 1-14.

DOI: https://doi.org/10.34238/tnu-jst.7733

Refbacks

There are currently no refbacks.



Remember me