Rapid-Learning Collaborative Pushing and Grasping via Deep Reinforcement Learning and Image Masking DOI Creative Commons
Chih-Yung Huang,

Guo Liang Su,

Yu-Hsiang Shao

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(19), P. 9018 - 9018

Published: Oct. 6, 2024

When multiple objects are positioned close together or stacked, pre-grasp operations such as pushing can be used to create space for the grasp, thereby improving grasping success rate. This study develops a model based on deep Q-learning network architecture and introduces fully convolutional accurately identify pixels in workspace image that correspond target locations exploration. In addition, this incorporates masking limit exploration area of robotic arm, ensuring agent consistently explores regions containing objects. approach effectively addresses sparse reward problem improves convergence rate model. Experimental results from both simulated real-world environments show proposed method accelerates learning effective strategies. is applied, task reaches 80% after 600 iterations. The time required reach 25% shorter when compared it not used. main finding direct integration technique with reinforcement (DRL) algorithm, which offers significant advancement arm control. Furthermore, shows substantially reduce training improve object innovation enables better adapt scenarios conventional DRL methods cannot handle, efficiency performance complex dynamic industrial applications.

Language: Английский

Student’s t-uniform mixture-based robust sparse coding model for sign language recognition from thermal images DOI
Saibal Ghosh,

Aninda Sundar Mondal,

Amitava Chatterjee

et al.

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 116619 - 116619

Published: Jan. 1, 2025

Language: Английский

Citations

1
HFNet: High-precision robotic grasp detection in unstructured environments using hierarchical RGB-D feature fusion and fine-grained pose alignment DOI
Ling Tong, Kun Qian

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 117775 - 117775

Published: May 1, 2025

Language: Английский

Citations

0
Rapid-Learning Collaborative Pushing and Grasping via Deep Reinforcement Learning and Image Masking DOI Creative Commons
Chih-Yung Huang,

Guo Liang Su,

Yu-Hsiang Shao

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(19), P. 9018 - 9018

Published: Oct. 6, 2024

When multiple objects are positioned close together or stacked, pre-grasp operations such as pushing can be used to create space for the grasp, thereby improving grasping success rate. This study develops a model based on deep Q-learning network architecture and introduces fully convolutional accurately identify pixels in workspace image that correspond target locations exploration. In addition, this incorporates masking limit exploration area of robotic arm, ensuring agent consistently explores regions containing objects. approach effectively addresses sparse reward problem improves convergence rate model. Experimental results from both simulated real-world environments show proposed method accelerates learning effective strategies. is applied, task reaches 80% after 600 iterations. The time required reach 25% shorter when compared it not used. main finding direct integration technique with reinforcement (DRL) algorithm, which offers significant advancement arm control. Furthermore, shows substantially reduce training improve object innovation enables better adapt scenarios conventional DRL methods cannot handle, efficiency performance complex dynamic industrial applications.

Language: Английский

Citations

0