Teaching robots behavior patterns by using reinforcement learning: How to raise pet robots with a remote control

Mans Ullerstam, Makoto Mizukawa

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

3 Citations (Scopus)

Abstract

The goal of this project was to show that complex behavior patterns can be learnt by a system based on reinforcement learning. The specific task was to make AIBO, the Sony robot dog, learn complex behavior patterns based on interactions between humans and AIBO. The reinforcement learning system is taught by remote control, used by the human and connected to AIBO. To remember the learnt behavior sequences, a short-term memory of prior actions is used by AIBO. This paper demonstrates that it is possible to learn behavior sequences and the relationship of cause and effect in complex environments. The paper also shows that the system works in a natural environment, based on the interaction between humans and AIBO, learning the rewards and the means to reach them in parallel. AIBO is also able to pick up new behaviors instantly by using a method we call 'Instant learning'. The paper presents the methods for implementing such a system.

Original languageEnglish
Title of host publicationProceedings of the SICE Annual Conference
Pages2251-2254
Number of pages4
Publication statusPublished - 2004
EventSICE Annual Conference 2004 - Sapporo
Duration: 2004 Aug 42004 Aug 6

Other

OtherSICE Annual Conference 2004
CitySapporo
Period04/8/404/8/6

Fingerprint

Reinforcement learning
Remote control
Teaching
Robots
Learning systems
Data storage equipment

Keywords

  • AIBO
  • Reinforcment learning
  • Remote control
  • User demonstration

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Ullerstam, M., & Mizukawa, M. (2004). Teaching robots behavior patterns by using reinforcement learning: How to raise pet robots with a remote control. In Proceedings of the SICE Annual Conference (pp. 2251-2254). [FPI-1-1]

Teaching robots behavior patterns by using reinforcement learning : How to raise pet robots with a remote control. / Ullerstam, Mans; Mizukawa, Makoto.

Proceedings of the SICE Annual Conference. 2004. p. 2251-2254 FPI-1-1.

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

Ullerstam, M & Mizukawa, M 2004, Teaching robots behavior patterns by using reinforcement learning: How to raise pet robots with a remote control. in Proceedings of the SICE Annual Conference., FPI-1-1, pp. 2251-2254, SICE Annual Conference 2004, Sapporo, 04/8/4.
Ullerstam M, Mizukawa M. Teaching robots behavior patterns by using reinforcement learning: How to raise pet robots with a remote control. In Proceedings of the SICE Annual Conference. 2004. p. 2251-2254. FPI-1-1
Ullerstam, Mans ; Mizukawa, Makoto. / Teaching robots behavior patterns by using reinforcement learning : How to raise pet robots with a remote control. Proceedings of the SICE Annual Conference. 2004. pp. 2251-2254
@inproceedings{03ea0c585b8043379c8f91df27492871,
title = "Teaching robots behavior patterns by using reinforcement learning: How to raise pet robots with a remote control",
abstract = "The goal of this project was to show that complex behavior patterns can be learnt by a system based on reinforcement learning. The specific task was to make AIBO, the Sony robot dog, learn complex behavior patterns based on interactions between humans and AIBO. The reinforcement learning system is taught by remote control, used by the human and connected to AIBO. To remember the learnt behavior sequences, a short-term memory of prior actions is used by AIBO. This paper demonstrates that it is possible to learn behavior sequences and the relationship of cause and effect in complex environments. The paper also shows that the system works in a natural environment, based on the interaction between humans and AIBO, learning the rewards and the means to reach them in parallel. AIBO is also able to pick up new behaviors instantly by using a method we call 'Instant learning'. The paper presents the methods for implementing such a system.",
keywords = "AIBO, Reinforcment learning, Remote control, User demonstration",
author = "Mans Ullerstam and Makoto Mizukawa",
year = "2004",
language = "English",
pages = "2251--2254",
booktitle = "Proceedings of the SICE Annual Conference",

}

TY - GEN

T1 - Teaching robots behavior patterns by using reinforcement learning

T2 - How to raise pet robots with a remote control

AU - Ullerstam, Mans

AU - Mizukawa, Makoto

PY - 2004

Y1 - 2004

N2 - The goal of this project was to show that complex behavior patterns can be learnt by a system based on reinforcement learning. The specific task was to make AIBO, the Sony robot dog, learn complex behavior patterns based on interactions between humans and AIBO. The reinforcement learning system is taught by remote control, used by the human and connected to AIBO. To remember the learnt behavior sequences, a short-term memory of prior actions is used by AIBO. This paper demonstrates that it is possible to learn behavior sequences and the relationship of cause and effect in complex environments. The paper also shows that the system works in a natural environment, based on the interaction between humans and AIBO, learning the rewards and the means to reach them in parallel. AIBO is also able to pick up new behaviors instantly by using a method we call 'Instant learning'. The paper presents the methods for implementing such a system.

AB - The goal of this project was to show that complex behavior patterns can be learnt by a system based on reinforcement learning. The specific task was to make AIBO, the Sony robot dog, learn complex behavior patterns based on interactions between humans and AIBO. The reinforcement learning system is taught by remote control, used by the human and connected to AIBO. To remember the learnt behavior sequences, a short-term memory of prior actions is used by AIBO. This paper demonstrates that it is possible to learn behavior sequences and the relationship of cause and effect in complex environments. The paper also shows that the system works in a natural environment, based on the interaction between humans and AIBO, learning the rewards and the means to reach them in parallel. AIBO is also able to pick up new behaviors instantly by using a method we call 'Instant learning'. The paper presents the methods for implementing such a system.

KW - AIBO

KW - Reinforcment learning

KW - Remote control

KW - User demonstration

UR - http://www.scopus.com/inward/record.url?scp=12744254963&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=12744254963&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:12744254963

SP - 2251

EP - 2254

BT - Proceedings of the SICE Annual Conference

ER -