{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,25]],"date-time":"2026-08-25T02:03:17Z","timestamp":1787623397753,"version":"build-2736575974"},"reference-count":38,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,11,7]],"date-time":"2023-11-07T00:00:00Z","timestamp":1699315200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Robot. AI"],"abstract":"<jats:p>Multi-robot cooperative control has been extensively studied using model-based distributed control methods. However, such control methods rely on sensing and perception modules in a sequential pipeline design, and the separation of perception and controls may cause processing latencies and compounding errors that affect control performance. End-to-end learning overcomes this limitation by implementing direct learning from onboard sensing data, with control commands output to the robots. Challenges exist in end-to-end learning for multi-robot cooperative control, and previous results are not scalable. We propose in this article a novel decentralized cooperative control method for multi-robot formations using deep neural networks, in which inter-robot communication is modeled by a graph neural network (GNN). Our method takes LiDAR sensor data as input, and the control policy is learned from demonstrations that are provided by an expert controller for decentralized formation control. Although it is trained with a fixed number of robots, the learned control policy is scalable. Evaluation in a robot simulator demonstrates the triangular formation behavior of multi-robot teams of different sizes under the learned control policy.<\/jats:p>","DOI":"10.3389\/frobt.2023.1285412","type":"journal-article","created":{"date-parts":[[2023,11,7]],"date-time":"2023-11-07T07:56:46Z","timestamp":1699343806000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":9,"title":["End-to-end decentralized formation control using a graph neural network-based learning method"],"prefix":"10.3389","volume":"10","author":[{"given":"Chao","family":"Jiang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xinchi","family":"Huang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yi","family":"Guo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2023,11,7]]},"reference":[{"key":"B1","first-page":"1741","article-title":"Learning transferable cooperative behavior in multi-agent teams","author":"Agarwal","year":"2020"},{"key":"B2","doi-asserted-by":"publisher","first-page":"90","DOI":"10.3389\/frobt.2019.00090","article-title":"Robust formation control for multiple underwater vehicles","volume":"6","author":"Bechlioulis","year":"2019","journal-title":"Front. Robotics AI"},{"key":"B3","first-page":"8772","article-title":"A framework for real-world multi-robot systems running decentralized gnn-based policies","author":"Blumenkamp","year":"2022"},{"key":"B4","first-page":"5501","article-title":"Distance-based formation control of a three-robot system","volume-title":"Chinese control and decision conference","author":"Chen","year":"2019"},{"key":"B5","doi-asserted-by":"publisher","first-page":"495","DOI":"10.9746\/jcmsi.10.495","article-title":"Coordinated control of multi-robot systems: a survey","volume":"10","author":"Cort\u00e9s","year":"2017","journal-title":"SICE J. Control, Meas. Syst. Integration"},{"key":"B6","doi-asserted-by":"publisher","first-page":"1546","DOI":"10.1109\/tro.2020.2994002","article-title":"Towards generalization in target-driven visual navigation by using deep reinforcement learning","volume":"36","author":"Devo","year":"2020","journal-title":"IEEE Trans. Robotics"},{"key":"B7","doi-asserted-by":"publisher","first-page":"1884","DOI":"10.1109\/lra.2019.2898117","article-title":"How fast is too fast? the role of perception latency in high-speed sense and avoid","volume":"4","author":"Falanga","year":"2019","journal-title":"IEEE Robotics Automation Lett."},{"key":"B8","doi-asserted-by":"crossref","DOI":"10.1609\/aaai.v32i1.11794","article-title":"Counterfactual multi-agent policy gradients","author":"Foerster","year":"2018"},{"key":"B9","first-page":"1146","article-title":"Stabilising experience replay for deep multi-agent reinforcement learning","author":"Foerster","year":"2017"},{"key":"B10","first-page":"2137","article-title":"Learning to communicate with deep multi-agent reinforcement learning","volume-title":"Advances in neural information processing systems","author":"Foerster","year":"2016"},{"key":"B11","doi-asserted-by":"publisher","first-page":"128","DOI":"10.1109\/msp.2020.3016143","article-title":"Graphs, convolutions, and neural networks: from graph filters to graph neural networks","volume":"37","author":"Gama","year":"2020","journal-title":"IEEE Signal Process. Mag."},{"key":"B12","doi-asserted-by":"publisher","first-page":"1034","DOI":"10.1109\/tsp.2018.2887403","article-title":"Convolutional neural network architectures for signals supported on graphs","volume":"67","author":"Gama","year":"2018","journal-title":"IEEE Trans. Signal Process."},{"key":"B13","doi-asserted-by":"publisher","first-page":"895","DOI":"10.1007\/s10462-021-09996-w","article-title":"Multi-agent deep reinforcement learning: a survey","author":"Gronauer","year":"2021","journal-title":"Artif. Intell. Rev."},{"key":"B14","doi-asserted-by":"crossref","DOI":"10.1002\/9781119216131","volume-title":"Distributed cooperative control: emerging applications","author":"Guo","year":"2017"},{"key":"B15","first-page":"66","article-title":"Cooperative multi-agent control using deep reinforcement learning","author":"Gupta","year":"2017"},{"key":"B16","first-page":"758","article-title":"Learning decentralized control policies for multi-robot formation","author":"Jiang","year":"2019"},{"key":"B17","doi-asserted-by":"publisher","first-page":"743","DOI":"10.1109\/lcsys.2020.3005441","article-title":"Multi-robot guided policy search for learning decentralized swarm control","volume":"5","author":"Jiang","year":"2020","journal-title":"IEEE Control Syst. Lett."},{"key":"B18","first-page":"7265","article-title":"Learning attentional communication for multi-agent cooperation","author":"Jiang","year":"2018"},{"key":"B19","doi-asserted-by":"publisher","first-page":"3121","DOI":"10.1109\/tmech.2021.3110660","article-title":"Distributed formation control of drones with onboard perception","volume":"27","author":"Kabore","year":"2021","journal-title":"IEEE\/ASME Trans. Mechatronics"},{"key":"B20","first-page":"5129","article-title":"Self-supervised deep reinforcement learning with generalized computation graphs for robot navigation","author":"Kahn","year":"2018"},{"key":"B21","first-page":"8801","article-title":"Decentralized global connectivity maintenance for multi-robot navigation: a reinforcement learning approach","author":"Li","year":"2022"},{"key":"B22","first-page":"11785","article-title":"Graph neural networks for decentralized multi-robot path planning","author":"Li","year":"2020"},{"key":"B23","doi-asserted-by":"publisher","first-page":"eabg5810","DOI":"10.1126\/scirobotics.abg5810","article-title":"Learning high-speed flight in the wild","volume":"6","author":"Loquercio","year":"2021","journal-title":"Sci. Robotics"},{"key":"B24","first-page":"6382","article-title":"Multi-agent actor-critic for mixed cooperative-competitive environments","author":"Lowe","year":"2017"},{"key":"B25","doi-asserted-by":"crossref","DOI":"10.1515\/9781400835355","volume-title":"Graph theoretic methods in multiagent networks","author":"Mesbahi","year":"2010"},{"key":"B26","doi-asserted-by":"crossref","first-page":"349","DOI":"10.1007\/978-3-030-45183-7_27","article-title":"An overview of gradient descent algorithm optimization in machine learning: application in the ophthalmology field","volume-title":"Smart applications and data analysis","author":"Mustapha","year":"2020"},{"key":"B27","doi-asserted-by":"publisher","first-page":"3","DOI":"10.3389\/frobt.2015.00003","article-title":"Dynamic coverage control in unicycle multi-robot networks under anisotropic sensing","volume":"2","author":"Panagou","year":"2015","journal-title":"Front. Robotics AI"},{"key":"B28","first-page":"627","article-title":"A reduction of imitation learning and structured prediction to no-regret online learning","author":"Ross","year":"2011"},{"key":"B29","doi-asserted-by":"publisher","first-page":"61","DOI":"10.1109\/tnn.2008.2005605","article-title":"The graph neural network model","volume":"20","author":"Scarselli","year":"2008","journal-title":"IEEE Trans. Neural Netw."},{"key":"B30","doi-asserted-by":"publisher","first-page":"345","DOI":"10.1023\/a:1008942012299","article-title":"Multiagent systems: a survey from a machine learning perspective","volume":"8","author":"Stone","year":"2000","journal-title":"Aut. Robots"},{"key":"B31","first-page":"2244","article-title":"Learning multiagent communication with backpropagation","volume":"29","author":"Sukhbaatar","year":"2016","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"B32","first-page":"671","article-title":"Learning decentralized controllers for robot swarms with graph neural networks","author":"Tolstaya","year":""},{"key":"B33","first-page":"8944","article-title":"Multi-robot coverage and exploration using spatial graph neural networks","author":"Tolstaya","year":""},{"key":"B34","doi-asserted-by":"publisher","first-page":"719","DOI":"10.1109\/tmech.2021.3056099","article-title":"Formation reconstruction and trajectory replanning for multi-uav patrol","volume":"26","author":"Wang","year":"2021","journal-title":"IEEE\/ASME Trans. Mechatronics"},{"key":"B35","doi-asserted-by":"publisher","first-page":"4509","DOI":"10.1109\/lra.2020.3002198","article-title":"Learning scheduling policies for multi-robot coordination with graph attention networks","volume":"5","author":"Wang","year":"2020","journal-title":"IEEE Robotics Automation Lett."},{"key":"B36","first-page":"1661","article-title":"Relative distributed formation and obstacle avoidance with multi-agent reinforcement learning","author":"Yan","year":"2022"},{"key":"B37","first-page":"2534","article-title":"Perception-aware receding horizon navigation for MAVs","author":"Zhang","year":"2018"},{"key":"B38","first-page":"4092","article-title":"Clone swarms: learning to predict and control multi-robot systems by imitation","author":"Zhou","year":"2019"}],"container-title":["Frontiers in Robotics and AI"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2023.1285412\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,11,7]],"date-time":"2023-11-07T07:56:54Z","timestamp":1699343814000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2023.1285412\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,11,7]]},"references-count":38,"alternative-id":["10.3389\/frobt.2023.1285412"],"URL":"https:\/\/doi.org\/10.3389\/frobt.2023.1285412","relation":{},"ISSN":["2296-9144"],"issn-type":[{"value":"2296-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,11,7]]},"article-number":"1285412"}}