Dynamic Mode Decomposition for perturbation estimation in human robot interaction

Erik Berger, Mark Sastuba, David Vogt, Bernhard Jung, Hani Ben Amor

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

6 Citations (Scopus)

Abstract

In many settings, e.g. physical human-robot interaction, robotic behavior must be made robust against more or less spontaneous application of external forces. Typically, this problem is tackled by means of special purpose force sensors which are, however, not available on many robotic platforms. In contrast, we propose a machine learning approach suitable for more common, although often noisy sensors. This machine learning approach makes use of Dynamic Mode Decomposition (DMD) which is able to extract the dynamics of a nonlinear system. It is therefore well suited to separate noise from regular oscillations in sensor readings during cyclic robot movements under different behavior configurations. We demonstrate the feasibility of our approach with an example where physical forces are exerted on a humanoid robot during walking. In a training phase, a snapshot based DMD model for behavior specific parameter configurations is learned. During task execution the robot must detect and estimate the external forces exerted by a human interaction partner. We compare the DMD-based approach to other interpolation schemes and show that the former outperforms the latter particularly in the presence of sensor noise. We conclude that DMD which has so far been mostly used in other fields of science, particularly fluid mechanics, is also a highly promising method for robotics.

Original languageEnglish (US)
Title of host publicationIEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages593-600
Number of pages8
ISBN (Print)9781479967636
DOIs
StatePublished - Oct 15 2014
Externally publishedYes
Event23rd IEEE International Symposium on Robot and Human Interactive Communication, IEEE RO-MAN 2014 - Edinburgh, United Kingdom
Duration: Aug 25 2014Aug 29 2014

Other

Other23rd IEEE International Symposium on Robot and Human Interactive Communication, IEEE RO-MAN 2014
CountryUnited Kingdom
CityEdinburgh
Period8/25/148/29/14

Fingerprint

Human robot interaction
Decomposition
Robotics
Sensors
Robots
Learning systems
Fluid mechanics
Nonlinear systems
Interpolation

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Human-Computer Interaction

Cite this

Berger, E., Sastuba, M., Vogt, D., Jung, B., & Ben Amor, H. (2014). Dynamic Mode Decomposition for perturbation estimation in human robot interaction. In IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions (pp. 593-600). [6926317] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ROMAN.2014.6926317

Dynamic Mode Decomposition for perturbation estimation in human robot interaction. / Berger, Erik; Sastuba, Mark; Vogt, David; Jung, Bernhard; Ben Amor, Hani.

IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions. Institute of Electrical and Electronics Engineers Inc., 2014. p. 593-600 6926317.

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

Berger, E, Sastuba, M, Vogt, D, Jung, B & Ben Amor, H 2014, Dynamic Mode Decomposition for perturbation estimation in human robot interaction. in IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions., 6926317, Institute of Electrical and Electronics Engineers Inc., pp. 593-600, 23rd IEEE International Symposium on Robot and Human Interactive Communication, IEEE RO-MAN 2014, Edinburgh, United Kingdom, 8/25/14. https://doi.org/10.1109/ROMAN.2014.6926317
Berger E, Sastuba M, Vogt D, Jung B, Ben Amor H. Dynamic Mode Decomposition for perturbation estimation in human robot interaction. In IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions. Institute of Electrical and Electronics Engineers Inc. 2014. p. 593-600. 6926317 https://doi.org/10.1109/ROMAN.2014.6926317
Berger, Erik ; Sastuba, Mark ; Vogt, David ; Jung, Bernhard ; Ben Amor, Hani. / Dynamic Mode Decomposition for perturbation estimation in human robot interaction. IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 593-600
@inproceedings{1d0a8b41f98b4273a80c0211166cb853,
title = "Dynamic Mode Decomposition for perturbation estimation in human robot interaction",
abstract = "In many settings, e.g. physical human-robot interaction, robotic behavior must be made robust against more or less spontaneous application of external forces. Typically, this problem is tackled by means of special purpose force sensors which are, however, not available on many robotic platforms. In contrast, we propose a machine learning approach suitable for more common, although often noisy sensors. This machine learning approach makes use of Dynamic Mode Decomposition (DMD) which is able to extract the dynamics of a nonlinear system. It is therefore well suited to separate noise from regular oscillations in sensor readings during cyclic robot movements under different behavior configurations. We demonstrate the feasibility of our approach with an example where physical forces are exerted on a humanoid robot during walking. In a training phase, a snapshot based DMD model for behavior specific parameter configurations is learned. During task execution the robot must detect and estimate the external forces exerted by a human interaction partner. We compare the DMD-based approach to other interpolation schemes and show that the former outperforms the latter particularly in the presence of sensor noise. We conclude that DMD which has so far been mostly used in other fields of science, particularly fluid mechanics, is also a highly promising method for robotics.",
author = "Erik Berger and Mark Sastuba and David Vogt and Bernhard Jung and {Ben Amor}, Hani",
year = "2014",
month = "10",
day = "15",
doi = "10.1109/ROMAN.2014.6926317",
language = "English (US)",
isbn = "9781479967636",
pages = "593--600",
booktitle = "IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Dynamic Mode Decomposition for perturbation estimation in human robot interaction

AU - Berger, Erik

AU - Sastuba, Mark

AU - Vogt, David

AU - Jung, Bernhard

AU - Ben Amor, Hani

PY - 2014/10/15

Y1 - 2014/10/15

N2 - In many settings, e.g. physical human-robot interaction, robotic behavior must be made robust against more or less spontaneous application of external forces. Typically, this problem is tackled by means of special purpose force sensors which are, however, not available on many robotic platforms. In contrast, we propose a machine learning approach suitable for more common, although often noisy sensors. This machine learning approach makes use of Dynamic Mode Decomposition (DMD) which is able to extract the dynamics of a nonlinear system. It is therefore well suited to separate noise from regular oscillations in sensor readings during cyclic robot movements under different behavior configurations. We demonstrate the feasibility of our approach with an example where physical forces are exerted on a humanoid robot during walking. In a training phase, a snapshot based DMD model for behavior specific parameter configurations is learned. During task execution the robot must detect and estimate the external forces exerted by a human interaction partner. We compare the DMD-based approach to other interpolation schemes and show that the former outperforms the latter particularly in the presence of sensor noise. We conclude that DMD which has so far been mostly used in other fields of science, particularly fluid mechanics, is also a highly promising method for robotics.

AB - In many settings, e.g. physical human-robot interaction, robotic behavior must be made robust against more or less spontaneous application of external forces. Typically, this problem is tackled by means of special purpose force sensors which are, however, not available on many robotic platforms. In contrast, we propose a machine learning approach suitable for more common, although often noisy sensors. This machine learning approach makes use of Dynamic Mode Decomposition (DMD) which is able to extract the dynamics of a nonlinear system. It is therefore well suited to separate noise from regular oscillations in sensor readings during cyclic robot movements under different behavior configurations. We demonstrate the feasibility of our approach with an example where physical forces are exerted on a humanoid robot during walking. In a training phase, a snapshot based DMD model for behavior specific parameter configurations is learned. During task execution the robot must detect and estimate the external forces exerted by a human interaction partner. We compare the DMD-based approach to other interpolation schemes and show that the former outperforms the latter particularly in the presence of sensor noise. We conclude that DMD which has so far been mostly used in other fields of science, particularly fluid mechanics, is also a highly promising method for robotics.

UR - http://www.scopus.com/inward/record.url?scp=84937565566&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937565566&partnerID=8YFLogxK

U2 - 10.1109/ROMAN.2014.6926317

DO - 10.1109/ROMAN.2014.6926317

M3 - Conference contribution

SN - 9781479967636

SP - 593

EP - 600

BT - IEEE RO-MAN 2014 - 23rd IEEE International Symposium on Robot and Human Interactive Communication: Human-Robot Co-Existence: Adaptive Interfaces and Systems for Daily Life, Therapy, Assistance and Socially Engaging Interactions

PB - Institute of Electrical and Electronics Engineers Inc.

ER -