Learning motion primitives of object manipulation using Mimesis Model

Bidan Huang, Joanna J. Bryson, Tetsunari Inamura

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

Abstract

In this paper, we present a system to learn manipulation motion primitives from human demonstration. This system, based on the statistical model “Mimesis Model”, provides an easy-to-use human-interface for learning manipulation motion primitives, as well as a natural language interface allowing human to modify and instruct robot motions. The human-demonstrated manipulation motion primitives are initially encoded by Hidden Markov Models (HMM). The models are then projected to a topological space where they are labeled, and their similarities are represented as their distances in the space. We then explore the unknown area in this space by interpolation between known models. New motion primitives are thus generated from the unknown area to meet the new manipulation scenarios. We demonstrate this system by learning bimanual grasping strategies. The implemented system successfully reproduces and generalizes the motion primitives in different grasping scenarios.
LanguageEnglish
Title of host publication2013 IEEE International Conference on Robotics and Biomimetics (ROBIO)
PublisherIEEE
Pages1144-1150
Number of pages7
DOIs
StatusPublished - 2013
Event2013 IEEE International Conference on Robotics and Biomimetics, ROBIO 2013 - Shenzhen, China
Duration: 12 Dec 201314 Dec 2013

Conference

Conference2013 IEEE International Conference on Robotics and Biomimetics, ROBIO 2013
CountryChina
CityShenzhen
Period12/12/1314/12/13

Fingerprint

Hidden Markov models
Interpolation
Demonstrations
Robots
Statistical Models

Cite this

Huang, B., Bryson, J. J., & Inamura, T. (2013). Learning motion primitives of object manipulation using Mimesis Model. In 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO) (pp. 1144-1150). IEEE. DOI: 10.1109/ROBIO.2013.6739618

Learning motion primitives of object manipulation using Mimesis Model. / Huang, Bidan; Bryson, Joanna J.; Inamura, Tetsunari.

2013 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, 2013. p. 1144-1150.

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

Huang, B, Bryson, JJ & Inamura, T 2013, Learning motion primitives of object manipulation using Mimesis Model. in 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, pp. 1144-1150, 2013 IEEE International Conference on Robotics and Biomimetics, ROBIO 2013, Shenzhen, China, 12/12/13. DOI: 10.1109/ROBIO.2013.6739618
Huang B, Bryson JJ, Inamura T. Learning motion primitives of object manipulation using Mimesis Model. In 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE. 2013. p. 1144-1150. Available from, DOI: 10.1109/ROBIO.2013.6739618
Huang, Bidan ; Bryson, Joanna J. ; Inamura, Tetsunari. / Learning motion primitives of object manipulation using Mimesis Model. 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, 2013. pp. 1144-1150
@inproceedings{b6eb9b2be6df451f813a75f0bfbcf7f1,
title = "Learning motion primitives of object manipulation using Mimesis Model",
abstract = "In this paper, we present a system to learn manipulation motion primitives from human demonstration. This system, based on the statistical model “Mimesis Model”, provides an easy-to-use human-interface for learning manipulation motion primitives, as well as a natural language interface allowing human to modify and instruct robot motions. The human-demonstrated manipulation motion primitives are initially encoded by Hidden Markov Models (HMM). The models are then projected to a topological space where they are labeled, and their similarities are represented as their distances in the space. We then explore the unknown area in this space by interpolation between known models. New motion primitives are thus generated from the unknown area to meet the new manipulation scenarios. We demonstrate this system by learning bimanual grasping strategies. The implemented system successfully reproduces and generalizes the motion primitives in different grasping scenarios.",
author = "Bidan Huang and Bryson, {Joanna J.} and Tetsunari Inamura",
year = "2013",
doi = "10.1109/ROBIO.2013.6739618",
language = "English",
pages = "1144--1150",
booktitle = "2013 IEEE International Conference on Robotics and Biomimetics (ROBIO)",
publisher = "IEEE",
address = "USA United States",

}

TY - GEN

T1 - Learning motion primitives of object manipulation using Mimesis Model

AU - Huang,Bidan

AU - Bryson,Joanna J.

AU - Inamura,Tetsunari

PY - 2013

Y1 - 2013

N2 - In this paper, we present a system to learn manipulation motion primitives from human demonstration. This system, based on the statistical model “Mimesis Model”, provides an easy-to-use human-interface for learning manipulation motion primitives, as well as a natural language interface allowing human to modify and instruct robot motions. The human-demonstrated manipulation motion primitives are initially encoded by Hidden Markov Models (HMM). The models are then projected to a topological space where they are labeled, and their similarities are represented as their distances in the space. We then explore the unknown area in this space by interpolation between known models. New motion primitives are thus generated from the unknown area to meet the new manipulation scenarios. We demonstrate this system by learning bimanual grasping strategies. The implemented system successfully reproduces and generalizes the motion primitives in different grasping scenarios.

AB - In this paper, we present a system to learn manipulation motion primitives from human demonstration. This system, based on the statistical model “Mimesis Model”, provides an easy-to-use human-interface for learning manipulation motion primitives, as well as a natural language interface allowing human to modify and instruct robot motions. The human-demonstrated manipulation motion primitives are initially encoded by Hidden Markov Models (HMM). The models are then projected to a topological space where they are labeled, and their similarities are represented as their distances in the space. We then explore the unknown area in this space by interpolation between known models. New motion primitives are thus generated from the unknown area to meet the new manipulation scenarios. We demonstrate this system by learning bimanual grasping strategies. The implemented system successfully reproduces and generalizes the motion primitives in different grasping scenarios.

UR - http://dx.doi.org/10.1109/ROBIO.2013.6739618

U2 - 10.1109/ROBIO.2013.6739618

DO - 10.1109/ROBIO.2013.6739618

M3 - Conference contribution

SP - 1144

EP - 1150

BT - 2013 IEEE International Conference on Robotics and Biomimetics (ROBIO)

PB - IEEE

ER -