File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: DeepMNavigate: Deep reinforced multi-robot navigation unifying local & global collision avoidance

TitleDeepMNavigate: Deep reinforced multi-robot navigation unifying local & global collision avoidance
Authors
KeywordsMotion and Path Planning
Collision Avoidance
Reinforcement learning
Robot sensing systems
Intelligent robots
Issue Date2020
PublisherInstitute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000393
Citation
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS): Consumer Robots and Our Future, Virtual Conference, Las Vegas, USA, 25 October 2020 - 24 January 2021, p. 6952-6959 How to Cite?
AbstractWe present a novel algorithm (DeepMNavigate) for global multi-agent navigation in dense scenarios using deep reinforcement learning (DRL). Our approach uses local and global information for each robot from motion information maps. We use a three-layer CNN that takes these maps as input to generate a suitable action to drive each robot to its goal position. Our approach is general, learns an optimal policy using a multi-scenario, multi-state training algorithm, and can directly handle raw sensor measurements for local observations. We demonstrate the performance on dense, complex benchmarks with narrow passages and environments with tens of agents. We highlight the algorithm’s benefits over prior learning methods and geometric decentralized algorithms in complex scenarios.
DescriptionTuCT19 Learning in Motion Planning - Paper TuCT19.6
Persistent Identifierhttp://hdl.handle.net/10722/285032
ISSN

 

DC FieldValueLanguage
dc.contributor.authorTan, Q-
dc.contributor.authorFan, T-
dc.contributor.authorPan, J-
dc.contributor.authorManocha, D-
dc.date.accessioned2020-08-07T09:05:51Z-
dc.date.available2020-08-07T09:05:51Z-
dc.date.issued2020-
dc.identifier.citationIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS): Consumer Robots and Our Future, Virtual Conference, Las Vegas, USA, 25 October 2020 - 24 January 2021, p. 6952-6959-
dc.identifier.issn2153-0858-
dc.identifier.urihttp://hdl.handle.net/10722/285032-
dc.descriptionTuCT19 Learning in Motion Planning - Paper TuCT19.6-
dc.description.abstractWe present a novel algorithm (DeepMNavigate) for global multi-agent navigation in dense scenarios using deep reinforcement learning (DRL). Our approach uses local and global information for each robot from motion information maps. We use a three-layer CNN that takes these maps as input to generate a suitable action to drive each robot to its goal position. Our approach is general, learns an optimal policy using a multi-scenario, multi-state training algorithm, and can directly handle raw sensor measurements for local observations. We demonstrate the performance on dense, complex benchmarks with narrow passages and environments with tens of agents. We highlight the algorithm’s benefits over prior learning methods and geometric decentralized algorithms in complex scenarios.-
dc.languageeng-
dc.publisherInstitute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000393-
dc.relation.ispartofIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Proceedings-
dc.rightsIEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Proceedings. Copyright © Institute of Electrical and Electronics Engineers.-
dc.rights©2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.-
dc.subjectMotion and Path Planning-
dc.subjectCollision Avoidance-
dc.subjectReinforcement learning-
dc.subjectRobot sensing systems-
dc.subjectIntelligent robots-
dc.titleDeepMNavigate: Deep reinforced multi-robot navigation unifying local & global collision avoidance-
dc.typeConference_Paper-
dc.identifier.emailPan, J: jpan@cs.hku.hk-
dc.identifier.authorityPan, J=rp01984-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/IROS45743.2020.9341805-
dc.identifier.hkuros312223-
dc.identifier.spage6952-
dc.identifier.epage6959-
dc.publisher.placeUnited States-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats