Abstract
There are some difficulties in applying traditional reinforcement learning algorithms to motion control tasks of robot. Because most algorithms are concerned with discrete actions and based on the assumption of complete observability of the state. This paper deals with these two problems by combining the reinforcement learning algorithm and CTRNN learning algorithm. We carried out an experiment on the pendulum swing-up task without rotational speed information. It is shown that the information about the rotational speed, which is considered as a hidden state, is estimated and encoded on the activation of a context neuron. As a result, this task is accomplished in several hundred trials using the proposed algorithm.
| Original language | English |
|---|---|
| Title of host publication | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Pages | 387-396 |
| Number of pages | 10 |
| Volume | 4232 LNCS |
| Publication status | Published - 2006 |
| Event | 13th International Conference on Neural Information Processing, ICONIP 2006 - Hong Kong Duration: 2006 Oct 3 → 2006 Oct 6 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 4232 LNCS |
| ISSN (Print) | 03029743 |
| ISSN (Electronic) | 16113349 |
Other
| Other | 13th International Conference on Neural Information Processing, ICONIP 2006 |
|---|---|
| City | Hong Kong |
| Period | 06/10/3 → 06/10/6 |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science
Cite this
Reinforcement learning algorithm with CTRNN in continuous action space. / Arie, Hiroaki; Namikawa, Jun; Ogata, Tetsuya; Tani, Jun; Sugano, Shigeki.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4232 LNCS 2006. p. 387-396 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4232 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Reinforcement learning algorithm with CTRNN in continuous action space
AU - Arie, Hiroaki
AU - Namikawa, Jun
AU - Ogata, Tetsuya
AU - Tani, Jun
AU - Sugano, Shigeki
PY - 2006
Y1 - 2006
N2 - There are some difficulties in applying traditional reinforcement learning algorithms to motion control tasks of robot. Because most algorithms are concerned with discrete actions and based on the assumption of complete observability of the state. This paper deals with these two problems by combining the reinforcement learning algorithm and CTRNN learning algorithm. We carried out an experiment on the pendulum swing-up task without rotational speed information. It is shown that the information about the rotational speed, which is considered as a hidden state, is estimated and encoded on the activation of a context neuron. As a result, this task is accomplished in several hundred trials using the proposed algorithm.
AB - There are some difficulties in applying traditional reinforcement learning algorithms to motion control tasks of robot. Because most algorithms are concerned with discrete actions and based on the assumption of complete observability of the state. This paper deals with these two problems by combining the reinforcement learning algorithm and CTRNN learning algorithm. We carried out an experiment on the pendulum swing-up task without rotational speed information. It is shown that the information about the rotational speed, which is considered as a hidden state, is estimated and encoded on the activation of a context neuron. As a result, this task is accomplished in several hundred trials using the proposed algorithm.
UR - http://www.scopus.com/inward/record.url?scp=33750590179&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33750590179&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:33750590179
SN - 3540464794
SN - 9783540464792
VL - 4232 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 387
EP - 396
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ER -