Motion recognition by combining HMM and reinforcement learning

Kazuhisa Hamamoto, Ken'ichi Morooka, Hiroshi Nagahashi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

It is difficult to give a robot all possible motions beforehand in a certain environment. Therefore, the robot needs to learn how to recognize others' motions and to generate its own motions autonomously for working well. These learning algorithms need an efficient way to make recognition and generation of motions work together, because they take many computing resources. This paper focuses on a generation-based recognition. Our system consists of recognition and generation modules. The former and latter are constructed from lefl-to-right Hidden Markov Models (HMM) and Reinforcement Learning (RL), respectively. When a HMM in recognition module does not work enough, the model parameters of HMM are re-estimated by using a state-value function of RL in generation module. The proposed method enables us to improve the reliability of the HMM.

Original languageEnglish
Title of host publication2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004
Pages5259-5264
Number of pages6
DOIs
Publication statusPublished - Dec 1 2004
Externally publishedYes
Event2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004 - The Hague, Netherlands
Duration: Oct 10 2004Oct 13 2004

Publication series

NameConference Proceedings - IEEE International Conference on Systems, Man and Cybernetics
Volume6
ISSN (Print)1062-922X

Other

Other2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004
CountryNetherlands
CityThe Hague
Period10/10/0410/13/04

Fingerprint

Reinforcement learning
Hidden Markov models
Robots
Learning algorithms

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Hamamoto, K., Morooka, K., & Nagahashi, H. (2004). Motion recognition by combining HMM and reinforcement learning. In 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004 (pp. 5259-5264). (Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics; Vol. 6). https://doi.org/10.1109/ICSMC.2004.1401029

Motion recognition by combining HMM and reinforcement learning. / Hamamoto, Kazuhisa; Morooka, Ken'ichi; Nagahashi, Hiroshi.

2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004. 2004. p. 5259-5264 (Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics; Vol. 6).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hamamoto, K, Morooka, K & Nagahashi, H 2004, Motion recognition by combining HMM and reinforcement learning. in 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004. Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics, vol. 6, pp. 5259-5264, 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004, The Hague, Netherlands, 10/10/04. https://doi.org/10.1109/ICSMC.2004.1401029
Hamamoto K, Morooka K, Nagahashi H. Motion recognition by combining HMM and reinforcement learning. In 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004. 2004. p. 5259-5264. (Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics). https://doi.org/10.1109/ICSMC.2004.1401029
Hamamoto, Kazuhisa ; Morooka, Ken'ichi ; Nagahashi, Hiroshi. / Motion recognition by combining HMM and reinforcement learning. 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004. 2004. pp. 5259-5264 (Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics).
@inproceedings{cbc7fa42c985430496f8cf27534c41bf,
title = "Motion recognition by combining HMM and reinforcement learning",
abstract = "It is difficult to give a robot all possible motions beforehand in a certain environment. Therefore, the robot needs to learn how to recognize others' motions and to generate its own motions autonomously for working well. These learning algorithms need an efficient way to make recognition and generation of motions work together, because they take many computing resources. This paper focuses on a generation-based recognition. Our system consists of recognition and generation modules. The former and latter are constructed from lefl-to-right Hidden Markov Models (HMM) and Reinforcement Learning (RL), respectively. When a HMM in recognition module does not work enough, the model parameters of HMM are re-estimated by using a state-value function of RL in generation module. The proposed method enables us to improve the reliability of the HMM.",
author = "Kazuhisa Hamamoto and Ken'ichi Morooka and Hiroshi Nagahashi",
year = "2004",
month = "12",
day = "1",
doi = "10.1109/ICSMC.2004.1401029",
language = "English",
isbn = "0780385667",
series = "Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics",
pages = "5259--5264",
booktitle = "2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004",

}

TY - GEN

T1 - Motion recognition by combining HMM and reinforcement learning

AU - Hamamoto, Kazuhisa

AU - Morooka, Ken'ichi

AU - Nagahashi, Hiroshi

PY - 2004/12/1

Y1 - 2004/12/1

N2 - It is difficult to give a robot all possible motions beforehand in a certain environment. Therefore, the robot needs to learn how to recognize others' motions and to generate its own motions autonomously for working well. These learning algorithms need an efficient way to make recognition and generation of motions work together, because they take many computing resources. This paper focuses on a generation-based recognition. Our system consists of recognition and generation modules. The former and latter are constructed from lefl-to-right Hidden Markov Models (HMM) and Reinforcement Learning (RL), respectively. When a HMM in recognition module does not work enough, the model parameters of HMM are re-estimated by using a state-value function of RL in generation module. The proposed method enables us to improve the reliability of the HMM.

AB - It is difficult to give a robot all possible motions beforehand in a certain environment. Therefore, the robot needs to learn how to recognize others' motions and to generate its own motions autonomously for working well. These learning algorithms need an efficient way to make recognition and generation of motions work together, because they take many computing resources. This paper focuses on a generation-based recognition. Our system consists of recognition and generation modules. The former and latter are constructed from lefl-to-right Hidden Markov Models (HMM) and Reinforcement Learning (RL), respectively. When a HMM in recognition module does not work enough, the model parameters of HMM are re-estimated by using a state-value function of RL in generation module. The proposed method enables us to improve the reliability of the HMM.

UR - http://www.scopus.com/inward/record.url?scp=15744366352&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=15744366352&partnerID=8YFLogxK

U2 - 10.1109/ICSMC.2004.1401029

DO - 10.1109/ICSMC.2004.1401029

M3 - Conference contribution

AN - SCOPUS:15744366352

SN - 0780385667

T3 - Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics

SP - 5259

EP - 5264

BT - 2004 IEEE International Conference on Systems, Man and Cybernetics, SMC 2004

ER -