GAPPO-ERW: an imitation reinforcement learning approach for AGV path planning


연구 분야: Artificial Intelligence



학회: The Journal of Supercomputing


초록

In the realm of Automated Guided Vehicle (AGV) path planning, this paper introduces an innovative Generative Adversarial Proximal Policy Optimization (GAPPO) approach, which integrates Generative Adversarial Imitation Learning (GAIL) with Proximal Policy Optimization (PPO). Additionally, an Expert-driven Adaptive Reward Weighting (ERW) strategy is incorporated to enhance the decision-making capabilities of the agent in complex environments. Through simulation validation in a virtual environment, the feasibility of this method in optimizing AGV systems has been demonstrated. Comparative results with other reinforcement learning techniques reveal that both GAPPO and GAPPO-ERW surpass traditional reinforcement learning methods in terms of path planning effectiveness and model training efficiency, showcasing their significant potential in enhancing AGV operational efficiency and flexibility.


Author Profile
Weiqiang Chen

Shenyang Institute of Computing Technology Chinese Academy of Sciences Nanping East Road No. 16 Shenyang 110168 People’s Republic of China

China
Author Profile
Xusheng Lin

School of Computer Science and Technology University of Chinese Academy of Sciences No. 19(A) Yuquan Road Beijing 100049 People’s Republic of China

Andorra
Author Profile
Zheng Zhou

Shenyang Institute of Computing Technology Chinese Academy of Sciences Nanping East Road No. 16 Shenyang 110168 People’s Republic of China

China

📄 논문 정보

발행 연도 2025년
인용수 0
출판 국가 Andorra, China
사이트 Springer
좋아요 수 0

연관 논문 목록 (305건)