Accommodating and assisting human partners in human-robot collaborative tasks through emotion understanding

Hope Diamantopoulos, Weitian Wang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Human-robot collaboration is inevitable in future smart manufacturing. In order to ensure safe human-robot collaboration, we develop a transfer learning-based approach to teach the robot to understand human emotions in smart manufacturing contexts. Our approach enables the robot to accurately understand human emotions and accommodate its human partner with corresponding assisting actions in collaborative tasks. This allows for improved collaboration-safety and enhanced ergonomics in human-robot partnerships. Experimental results and evaluations demonstrate that the robot can precisely understand human emotions in real time and effectively assist its human partner in real-world co-assembly tasks. Future work for this study is also discussed.

Original languageEnglish
Title of host publication2021 12th International Conference on Mechanical and Aerospace Engineering, ICMAE 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages523-528
Number of pages6
ISBN (Electronic)9781665433211
DOIs
StatePublished - 16 Jul 2021
Event12th International Conference on Mechanical and Aerospace Engineering, ICMAE 2021 - Virtual, Athens, Greece
Duration: 16 Jul 202119 Jul 2021

Publication series

Name2021 12th International Conference on Mechanical and Aerospace Engineering, ICMAE 2021

Conference

Conference12th International Conference on Mechanical and Aerospace Engineering, ICMAE 2021
Country/TerritoryGreece
CityVirtual, Athens
Period16/07/2119/07/21

Keywords

  • Human-robot interaction
  • Intelligent control
  • Machine learning
  • Robotics
  • Vision system

Fingerprint

Dive into the research topics of 'Accommodating and assisting human partners in human-robot collaborative tasks through emotion understanding'. Together they form a unique fingerprint.

Cite this