Active sensing based dynamical object feature extraction

Shun Nishide, Tetsuya Ogata, Ryunosuke Yokoya, Jun Tani, Kazunori Komatani, Hiroshi G. Okuno

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

3 Citations (Scopus)

Abstract

This paper presents a method to autonomously extract object features that describe their dynamics from active sensing experiences. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for extracting object features that describe the object motions. The two modules are simultaneously trained using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing, and rolling motions. The results have shown that the model is capable of extracting features that distinguish the characteristics of object dynamics.

Original languageEnglish
Title of host publication2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS
Pages1-7
Number of pages7
DOIs
Publication statusPublished - 2008
Externally publishedYes
Event2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS - Nice
Duration: 2008 Sep 222008 Sep 26

Other

Other2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS
CityNice
Period08/9/2208/9/26

Fingerprint

Feature extraction
Recurrent neural networks
Robots
Neural networks
Experiments

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Nishide, S., Ogata, T., Yokoya, R., Tani, J., Komatani, K., & Okuno, H. G. (2008). Active sensing based dynamical object feature extraction. In 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS (pp. 1-7). [4650794] https://doi.org/10.1109/IROS.2008.4650794

Active sensing based dynamical object feature extraction. / Nishide, Shun; Ogata, Tetsuya; Yokoya, Ryunosuke; Tani, Jun; Komatani, Kazunori; Okuno, Hiroshi G.

2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS. 2008. p. 1-7 4650794.

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

Nishide, S, Ogata, T, Yokoya, R, Tani, J, Komatani, K & Okuno, HG 2008, Active sensing based dynamical object feature extraction. in 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS., 4650794, pp. 1-7, 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, Nice, 08/9/22. https://doi.org/10.1109/IROS.2008.4650794
Nishide S, Ogata T, Yokoya R, Tani J, Komatani K, Okuno HG. Active sensing based dynamical object feature extraction. In 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS. 2008. p. 1-7. 4650794 https://doi.org/10.1109/IROS.2008.4650794
Nishide, Shun ; Ogata, Tetsuya ; Yokoya, Ryunosuke ; Tani, Jun ; Komatani, Kazunori ; Okuno, Hiroshi G. / Active sensing based dynamical object feature extraction. 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS. 2008. pp. 1-7
@inproceedings{56ca6113ceff45e1bd5b76ff1ed64dbf,
title = "Active sensing based dynamical object feature extraction",
abstract = "This paper presents a method to autonomously extract object features that describe their dynamics from active sensing experiences. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for extracting object features that describe the object motions. The two modules are simultaneously trained using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing, and rolling motions. The results have shown that the model is capable of extracting features that distinguish the characteristics of object dynamics.",
author = "Shun Nishide and Tetsuya Ogata and Ryunosuke Yokoya and Jun Tani and Kazunori Komatani and Okuno, {Hiroshi G.}",
year = "2008",
doi = "10.1109/IROS.2008.4650794",
language = "English",
isbn = "9781424420582",
pages = "1--7",
booktitle = "2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS",

}

TY - GEN

T1 - Active sensing based dynamical object feature extraction

AU - Nishide, Shun

AU - Ogata, Tetsuya

AU - Yokoya, Ryunosuke

AU - Tani, Jun

AU - Komatani, Kazunori

AU - Okuno, Hiroshi G.

PY - 2008

Y1 - 2008

N2 - This paper presents a method to autonomously extract object features that describe their dynamics from active sensing experiences. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for extracting object features that describe the object motions. The two modules are simultaneously trained using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing, and rolling motions. The results have shown that the model is capable of extracting features that distinguish the characteristics of object dynamics.

AB - This paper presents a method to autonomously extract object features that describe their dynamics from active sensing experiences. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for extracting object features that describe the object motions. The two modules are simultaneously trained using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing, and rolling motions. The results have shown that the model is capable of extracting features that distinguish the characteristics of object dynamics.

UR - http://www.scopus.com/inward/record.url?scp=69549110096&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=69549110096&partnerID=8YFLogxK

U2 - 10.1109/IROS.2008.4650794

DO - 10.1109/IROS.2008.4650794

M3 - Conference contribution

SN - 9781424420582

SP - 1

EP - 7

BT - 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS

ER -