Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network
Proper interaction is a crucial aspect of team collaborations for successfully achieving a common goal. In recent times, more technically advanced robots have been introduced into the industrial environments sharing the same workspace as other robots and humans which causes the need for human-robot...
Main Authors: | , |
---|---|
Format: | Others |
Language: | English |
Published: |
Mälardalens högskola, Akademin för innovation, design och teknik
2020
|
Subjects: | |
Online Access: | http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-50918 |
id |
ndltd-UPSALLA1-oai-DiVA.org-mdh-50918 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-UPSALLA1-oai-DiVA.org-mdh-509182020-10-06T05:28:02ZHuman-Robot Interaction Using Reinforcement Learning and Convolutional Neural NetworkengKhan, YousufOtalvaro, EdierMälardalens högskola, Akademin för innovation, design och teknikMälardalens högskola, Akademin för innovation, design och teknik2020RoboticsRobotteknik och automationProper interaction is a crucial aspect of team collaborations for successfully achieving a common goal. In recent times, more technically advanced robots have been introduced into the industrial environments sharing the same workspace as other robots and humans which causes the need for human-robot interaction (HRI) to be greater than ever before. The purpose of this study is to enable a HRI by teaching a robot to classify different human facial expressions as either positive or negative using a convolutional neural network and respond to each of them with the help of the reinforcement learning algorithm Q-learning.The simulation showed that the robot could accurately classify and react to the facial expressions under the instructions given by the Q-learning algorithm. The simulated results proved to be consistent in every conducted experiment having low variances. These results are promising for future research to allow for the study to be conducted in real-life environments. Student thesisinfo:eu-repo/semantics/bachelorThesistexthttp://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-50918application/pdfinfo:eu-repo/semantics/openAccess |
collection |
NDLTD |
language |
English |
format |
Others
|
sources |
NDLTD |
topic |
Robotics Robotteknik och automation |
spellingShingle |
Robotics Robotteknik och automation Khan, Yousuf Otalvaro, Edier Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
description |
Proper interaction is a crucial aspect of team collaborations for successfully achieving a common goal. In recent times, more technically advanced robots have been introduced into the industrial environments sharing the same workspace as other robots and humans which causes the need for human-robot interaction (HRI) to be greater than ever before. The purpose of this study is to enable a HRI by teaching a robot to classify different human facial expressions as either positive or negative using a convolutional neural network and respond to each of them with the help of the reinforcement learning algorithm Q-learning.The simulation showed that the robot could accurately classify and react to the facial expressions under the instructions given by the Q-learning algorithm. The simulated results proved to be consistent in every conducted experiment having low variances. These results are promising for future research to allow for the study to be conducted in real-life environments. |
author |
Khan, Yousuf Otalvaro, Edier |
author_facet |
Khan, Yousuf Otalvaro, Edier |
author_sort |
Khan, Yousuf |
title |
Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
title_short |
Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
title_full |
Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
title_fullStr |
Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
title_full_unstemmed |
Human-Robot Interaction Using Reinforcement Learning and Convolutional Neural Network |
title_sort |
human-robot interaction using reinforcement learning and convolutional neural network |
publisher |
Mälardalens högskola, Akademin för innovation, design och teknik |
publishDate |
2020 |
url |
http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-50918 |
work_keys_str_mv |
AT khanyousuf humanrobotinteractionusingreinforcementlearningandconvolutionalneuralnetwork AT otalvaroedier humanrobotinteractionusingreinforcementlearningandconvolutionalneuralnetwork |
_version_ |
1719350745682149376 |