Print Email Facebook Twitter DACOOP-A Title DACOOP-A: Decentralized Adaptive Cooperative Pursuit via Attention Author Zhang, Zheng (Sun Yat-sen University) Zhang, Dengyu (Sun Yat-sen University) Zhang, Qingrui (Sun Yat-sen University) Pan, W. (TU Delft Robot Dynamics; The University of Manchester) Hu, Tianjiang (Sun Yat-sen University) Date 2024 Abstract Integrating rule-based policies into reinforcement learning promises to improve data efficiency and generalization in cooperative pursuit problems. However, most implementations do not properly distinguish the influence of neighboring robots in observation embedding or inter-robot interaction rules, leading to information loss and inefficient cooperation. This letter proposes a cooperative pursuit algorithm named Decentralized Adaptive COOperative Pursuit via Attention (DACOOP-A) by empowering reinforcement learning with artificial potential field and attention mechanisms. An attention-based framework is developed to emphasize important neighbors by concurrently integrating the learned attention scores into observation embedding and inter-robot interaction rules. A KL divergence regularization is introduced to alleviate the resultant learning stability issue. Improvements in data efficiency and generalization are demonstrated through numerical simulations. Extensive quantitative analyses are performed to illustrate the advantages of the proposed modules. Real-world experiments are performed to justify the feasibility of DACOOP-A in physical systems. Subject Attention mechanismcooperative pursuitmulti-robot systemsreinforcement learning To reference this document use: http://resolver.tudelft.nl/uuid:ae4a2a8a-7f58-4c06-ac0b-202b8e02c0a8 DOI https://doi.org/10.1109/LRA.2023.3331886 Embargo date 2024-05-10 ISSN 2377-3766 Source IEEE Robotics and Automation Letters, 9 (6), 5504-5511 Bibliographical note Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. Part of collection Institutional Repository Document type journal article Rights © 2024 Zheng Zhang, Dengyu Zhang, Qingrui Zhang, W. Pan, Tianjiang Hu Files PDF DACOOP-A_Decentralized_Ad ... ention.pdf 3.66 MB Close viewer /islandora/object/uuid:ae4a2a8a-7f58-4c06-ac0b-202b8e02c0a8/datastream/OBJ/view