Multimodal control for human-robot cooperation

For intuitive human-robot collaboration, the robot must quickly adapt to the human behavior. To this end, we propose a multimodal sensor-based control framework, enabling a robot to recognize human intention, and consequently adapt its control strategy. Our approach is marker-less, relies on a Kinect and on an on-board camera, and is based on a unified task formalism. Moreover, we validate it in a mock-up industrial scenario, where human and robot must collaborate to insert screws in a flank (see [CPM2013]).

[CPM2013] A. Cherubini, R. Passama, A. Meline, and P. Fraisse. Multimodal control for human-robot cooperation. In Ieee/rsj international conference on intelligent robots and systems (iros), nov 2013.
[Bibtex]
@INPROCEEDINGS{CPM2013,
author={Cherubini, A. and Passama, R. and Meline, A. and Fraisse, P.},
booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
title={Multimodal control for human-robot cooperation},
year={2013},
month={nov},
volume={},
number={},
pages={},
keywords={iros, Human-Robot Interaction},
doi={},
ISSN={}}



Posted on by fraisse | Comments Off

Online Identification and Visualization of the Statically Equivalent Serial Chain via constrained Kalman filter

A human’s center of mass (CoM) trajectory is useful to evaluate the dynamic stability during daily life activities such as walking and standing up. To estimate the subject specific CoM position in the home environment, we make use of a statically equivalent serial chain (SESC) developed with a portable measurement system. In this paper we implement a constrained Kalman filter to achieve an online parameter estimation of the SESC parameters while accounting for the human body bilateral symmetry. This results in constraining SESC parameters to be consistent with the human skeletal model used. Kinect is used as a markerless motion capture system for measuring limb orientations while the Wii board is used to measure the subject’s center of pressure (CoP) during the identification phase. CoP measurements and Kinect data were recorded for five able-bodied subjects. The recorded data was then given to the proposed recursive algorithm to identify the parameters of the SESC online (see [GHF2013]).

[GHF2013] A. Gonzalez, M. Hayashibe, and P. Fraisse. Online identification and visualization of the statically equivalent serial chain via constrained kalman filter. In Ieee international conference on robotics and automation (icra), oct 2013.
[Bibtex]
@INPROCEEDINGS{GHF2013,
author={Gonzalez, A. and Hayashibe, M. and Fraisse, P.},
booktitle={IEEE International Conference on Robotics and Automation (ICRA)},
title={Online Identification and Visualization of the Statically Equivalent Serial Chain via constrained Kalman filter},
year={2013},
month={oct},
volume={},
number={},
pages={},
keywords={iros, Human modeling},
doi={},
ISSN={}}


Posted on by fraisse | Comments Off

IROS 2012: Estimation of the Center of Mass with Kinect and Wii balance board
  • [PDF] A. Gonzalez, M. Hayashibe, and P. Fraisse. Estimation of the center of mass with kinect and wii balance board. In Ieee/rsj international conference on intelligent robots and systems (iros), oct 2012.
    [Bibtex]
    @INPROCEEDINGS{2012-iros-ag,
    author={Gonzalez, A. and Hayashibe, M. and Fraisse, P.},
    booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    title={Estimation of the Center of Mass with Kinect and Wii balance board},
    year={2012},
    month={oct},
    volume={},
    number={},
    pages={},
    keywords={iros, Human modeling},
    doi={},
    ISSN={}}

Center of mass (CoM) trajectory is important during standing and walking since it can be used as an index for stability and fall prediction. Unfortunately current methods for CoM estimation require the use of specialized equipment (such as motion capture and force platforms) in controlled environments. This paper aims at applying the statically equivalent serial chain (SESC) method to obtain CoM position using widely available and portable hardware; a Microsoft’s Kinect and a Nintendo’s Wii balance board. During identification, CoM is approximated by CoP measurements and the virtual chain is created for able-bodied subjects. The result demostrates that the SESC method can be applied outside the laboratory environment using a Kinect. Cross-validation of the identified model was performed to evaluate the accuracy of the method. Results obtained of five subjects are shown and discussed.


Posted on by fraisse | Comments Off

IROS 2011: Interactive Manipulation Between a Human and a Humanoid: When Robots Control Human Arm Motion

  • [PDF] [DOI] B. V. Adorno, A. P. L. Bo, and P. Fraisse. Interactive manipulation between a human and a humanoid: when robots control human arm motion. In Intelligent robots and systems (iros), 2011 ieee/rsj international conference on, pages 4658-4663, sept 2011.
    [Bibtex]
    @INPROCEEDINGS{2011-iros-bva,
    author={Adorno, B. V. and Bo, A. P. L. and Fraisse, P.},
    booktitle={Intelligent Robots and Systems (IROS), 2011 IEEE/RSJ International Conference on},
    title={Interactive manipulation between a human and a humanoid: When robots control human arm motion},
    year={2011},
    month={sept},
    volume={},
    number={},
    pages={4658 -4663},
    keywords={iros},
    doi={10.1109/IROS.2011.6094869},
    ISSN={2153-0858}}

see IEEE SPECTRUM

In this paper we present a novel approach in human/robot collaboration, where the robot controls not only its arm but also the human’s by means of functional electrical stimulation (FES). The task is described by using the cooperative dual task-space approach, providing a considerable degree of invariance with respect to the morphology of the agents involved. Experimental results in a “ball in the hoop” task using healthy blindfolded subjects show the validity of the approach and encourages further experiments with impaired subjects, for instance hemiplegic or quadriplegic patients.


Posted on by fraisse | Comments Off

 Planning and Fast Replanning Safe Motions for Humanoid Robots

  • [PDF] [DOI] S. Lengagne, N. Ramdani, and P. Fraisse. Planning and fast replanning safe motions for humanoid robots. Ieee transactions on robotics, 27(6):1095-1106, February 2011.
    [Bibtex]
    @article{a2011-tro-sl-al,
    Author = {Lengagne, S. and Ramdani, N and Fraisse, P.},
    Journal = {IEEE Transactions on Robotics},
    Month = {February},
    Number = {6},
    Pages = {1095-1106},
    Title = {Planning and Fast Replanning Safe Motions for Humanoid Robots},
    Volume = {27},
    Year = {2011},
    keywords={publier},
    doi={10.1109/TRO.2011.2162998},
    }

This paper introduces effective numerical methods for the planning and fast replanning of safe motions to ensure the safety, balance and integrity of humanoid robots over the whole motion duration. Our safe methods do not depend nor are connected to any type of modelling or constraints. To plan safe motions, certain constraints have to be satisfied over a continuous interval of time. Classical methods revert to time-grid discretization, which can be risky for the robot. We introduce a hybrid method for planning safe motions, which combines a classical unsafe method with a verification step that checks constraint violation and computes excess using interval analysis. When the robot meets unexpected situations, it has to replan a new motion, which is often too time-consuming. Hence, we introduce a new method for rapidly replanning safe motions, i.e., in less than 2s CPU time. It computes off-line feasible subsets in the vicinity of safe motions and finds on-line a solution in these subsets without actually computing again the nonlinear constraints. Our methods are validated using the HOAP-3 robot, where the motions are run without any balance controller.


Posted on by fraisse | Comments Off