面向实时分布式多机器人任务分配的图强化学习框架

Q3 Earth and Planetary Sciences
Dian Zhang, Peng Dong, Pai Peng, Yubo Dong
{"title":"面向实时分布式多机器人任务分配的图强化学习框架","authors":"Dian Zhang,&nbsp;Peng Dong,&nbsp;Pai Peng,&nbsp;Yubo Dong","doi":"10.1007/s42401-024-00334-w","DOIUrl":null,"url":null,"abstract":"<div><p>Dynamic multi-robot task allocation (MRTA) requires real-time responsiveness and adaptability to rapidly changing conditions. Existing methods, primarily based on static data and centralized architectures, often fail in dynamic environments that require decentralized, context-aware decisions. To address these challenges, this paper proposes a novel graph reinforcement learning (GRL) architecture, named Spatial-Temporal Fusing Reinforcement Learning (STFRL), to address real-time distributed target allocation problems in search and rescue scenarios. The proposed policy network includes an encoder, which employs a Temporal-Spatial Fusing Encoder (TSFE) to extract input features and a decoder uses multi-head attention (MHA) to perform distributed allocation based on the encoder’s output and context. The policy network is trained with the REINFORCE algorithm. Experimental comparisons with state-of-the-art baselines demonstrate that STFRL achieves superior performance in path cost, inference speed, and scalability, highlighting its robustness and efficiency in complex, dynamic environments.</p></div>","PeriodicalId":36309,"journal":{"name":"Aerospace Systems","volume":"8 1","pages":"105 - 116"},"PeriodicalIF":0.0000,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s42401-024-00334-w.pdf","citationCount":"0","resultStr":"{\"title\":\"A graph reinforcement learning framework for real-time distributed multi-robot task allocation\",\"authors\":\"Dian Zhang,&nbsp;Peng Dong,&nbsp;Pai Peng,&nbsp;Yubo Dong\",\"doi\":\"10.1007/s42401-024-00334-w\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Dynamic multi-robot task allocation (MRTA) requires real-time responsiveness and adaptability to rapidly changing conditions. Existing methods, primarily based on static data and centralized architectures, often fail in dynamic environments that require decentralized, context-aware decisions. To address these challenges, this paper proposes a novel graph reinforcement learning (GRL) architecture, named Spatial-Temporal Fusing Reinforcement Learning (STFRL), to address real-time distributed target allocation problems in search and rescue scenarios. The proposed policy network includes an encoder, which employs a Temporal-Spatial Fusing Encoder (TSFE) to extract input features and a decoder uses multi-head attention (MHA) to perform distributed allocation based on the encoder’s output and context. The policy network is trained with the REINFORCE algorithm. Experimental comparisons with state-of-the-art baselines demonstrate that STFRL achieves superior performance in path cost, inference speed, and scalability, highlighting its robustness and efficiency in complex, dynamic environments.</p></div>\",\"PeriodicalId\":36309,\"journal\":{\"name\":\"Aerospace Systems\",\"volume\":\"8 1\",\"pages\":\"105 - 116\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2025-01-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://link.springer.com/content/pdf/10.1007/s42401-024-00334-w.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Aerospace Systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://link.springer.com/article/10.1007/s42401-024-00334-w\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"Earth and Planetary Sciences\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Aerospace Systems","FirstCategoryId":"1085","ListUrlMain":"https://link.springer.com/article/10.1007/s42401-024-00334-w","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"Earth and Planetary Sciences","Score":null,"Total":0}
引用次数: 0

摘要

动态多机器人任务分配(MRTA)需要实时响应和适应快速变化的条件。现有的方法主要基于静态数据和集中式架构,在需要分散的、上下文感知的决策的动态环境中往往会失败。为了解决这些挑战,本文提出了一种新的图强化学习(GRL)架构,称为时空融合强化学习(STFRL),以解决搜索和救援场景中的实时分布式目标分配问题。所提出的策略网络包括一个编码器,该编码器采用时空融合编码器(TSFE)提取输入特征,解码器使用多头注意(MHA)根据编码器的输出和上下文进行分布式分配。采用强化算法对策略网络进行训练。与最先进基线的实验比较表明,STFRL在路径成本、推理速度和可扩展性方面具有优越的性能,突出了其在复杂动态环境中的鲁棒性和效率。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
A graph reinforcement learning framework for real-time distributed multi-robot task allocation

Dynamic multi-robot task allocation (MRTA) requires real-time responsiveness and adaptability to rapidly changing conditions. Existing methods, primarily based on static data and centralized architectures, often fail in dynamic environments that require decentralized, context-aware decisions. To address these challenges, this paper proposes a novel graph reinforcement learning (GRL) architecture, named Spatial-Temporal Fusing Reinforcement Learning (STFRL), to address real-time distributed target allocation problems in search and rescue scenarios. The proposed policy network includes an encoder, which employs a Temporal-Spatial Fusing Encoder (TSFE) to extract input features and a decoder uses multi-head attention (MHA) to perform distributed allocation based on the encoder’s output and context. The policy network is trained with the REINFORCE algorithm. Experimental comparisons with state-of-the-art baselines demonstrate that STFRL achieves superior performance in path cost, inference speed, and scalability, highlighting its robustness and efficiency in complex, dynamic environments.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Aerospace Systems
Aerospace Systems Social Sciences-Social Sciences (miscellaneous)
CiteScore
1.80
自引率
0.00%
发文量
53
期刊介绍: Aerospace Systems provides an international, peer-reviewed forum which focuses on system-level research and development regarding aeronautics and astronautics. The journal emphasizes the unique role and increasing importance of informatics on aerospace. It fills a gap in current publishing coverage from outer space vehicles to atmospheric vehicles by highlighting interdisciplinary science, technology and engineering. Potential topics include, but are not limited to: Trans-space vehicle systems design and integration Air vehicle systems Space vehicle systems Near-space vehicle systems Aerospace robotics and unmanned system Communication, navigation and surveillance Aerodynamics and aircraft design Dynamics and control Aerospace propulsion Avionics system Opto-electronic system Air traffic management Earth observation Deep space exploration Bionic micro-aircraft/spacecraft Intelligent sensing and Information fusion
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信