DQN Reinforcement Learning: The Robot’s Optimum Path Navigation in Dynamic Environments for Smart Factory 


Vol. 44,  No. 12, pp. 2269-2279, Dec.  2019
10.7840/kics.2019.44.12.2269


PDF Full-Text
  Abstract

Most of the robots used in smart factory are controlled by users or are commercially available with line-tracer techniques. These robots are used to carry out accessories or repair by moving them to a designated position or to a position where alarms are activated. However, there is a problem with these methods that it is difficult to apply the existing algorithms due to frequently changing process line layout and a position where alarms are activated. To solve these problems, this paper proposes a reinforcement learning model that can be actively and flexibly dealt with the change of situation. The proposed model is a model using Deep Q-Network, one of the enhanced learning algorithms, that receives images of the internal structure and derives the optimal path of movement for the current location of robots. This method allows the establishment of a system that can be proactive and flexible in responding to the changes in circumstances. In addition, the system will greatly contribute to energy savings and productivity gains. The simulation results showed that even in new unlearned environments it was found that the proposed method effectively elicited the optimal path taking into account the structure and conditions inside the plant.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

K. Park, J. Park, W. Yun, S. Yoo, "DQN Reinforcement Learning: The Robot’s Optimum Path Navigation in Dynamic Environments for Smart Factory," The Journal of Korean Institute of Communications and Information Sciences, vol. 44, no. 12, pp. 2269-2279, 2019. DOI: 10.7840/kics.2019.44.12.2269.

[ACM Style]

Kwang-Seok Park, Jin-Man Park, Wan-Kyu Yun, and Sang-Jo Yoo. 2019. DQN Reinforcement Learning: The Robot’s Optimum Path Navigation in Dynamic Environments for Smart Factory. The Journal of Korean Institute of Communications and Information Sciences, 44, 12, (2019), 2269-2279. DOI: 10.7840/kics.2019.44.12.2269.

[KICS Style]

Kwang-Seok Park, Jin-Man Park, Wan-Kyu Yun, Sang-Jo Yoo, "DQN Reinforcement Learning: The Robot’s Optimum Path Navigation in Dynamic Environments for Smart Factory," The Journal of Korean Institute of Communications and Information Sciences, vol. 44, no. 12, pp. 2269-2279, 12. 2019. (https://doi.org/10.7840/kics.2019.44.12.2269)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880