Retour au site du LAAS-CNRS

Laboratoire d’analyse et d’architecture des systèmes

Publications de l'équipe gepetto

Choisir la langue : FR | EN

357documents trouvés

17052
01/05/2017

Adaptive synthesis of dynamically feasible full-body movements for the humanoid robot HRP-2 by flexible combination of learned dynamic movement primitives

A.MUKOVSKIY, C.VASSALLO, M.NAVEAU, O.STASSE, P.SOUERES, M.GIESE

Tubingen, GEPETTO

Revue Scientifique : Robotics and Autonomous Systems, Vol.91, pp.270-283, Mai 2017 , N° 17052

Lien : https://hal.archives-ouvertes.fr/hal-01484935

Diffusable

Plus d'informations

Abstract

Skilled human full-body movements are often planned in a highly predictive manner. For example, during walking while reaching towards a goal object results in steps and body postures are adapted to the goal position already multiple steps before the goal contact. The realization of such highly predictive behaviors for humanoid robots is a challenge because standard approaches, such as optimal control, result in computation times that are prohibitive for the predictive control of complex coordinated full-body movements over multiple steps. We devised a new architecture that combines the online-planning of complex coordinated full-body movements, based on the flexible combination of learned dynamic movement primitives, with a Walking Pattern Generator (WPG), based on Model Predictive Control (MPC), which generates dynamically feasible locomotion of the humanoid robot HRP-2. A dynamic filter corrects the Zero Moment Point (ZMP) trajectories in order to guarantee the dynamic feasibility of the executed behavior taking into account the upper-body movements, at the same time ensuring an accurate approximation of the planned motion trajectories. We demonstrate the high flexibility of the chosen movement planning approach, and the accuracy and feasibility of the generated motion. In addition, we show that a na¨ıvena¨ıve approach, which generates adaptive motion by using machine learning methods by the interpolation between feasible training motion examples fails to guarantee the stability and dynamic feasibility of the generated behaviors.

139327
17055
22/03/2017

TALOS: A new humanoid research platform targeted for industrial applications

O.STASSE, T.FLAYOLS, R.BUDHIRAJA, K.GIRAUD ESCLASSE, J.CARPENTIER, A.DEL PRETE, P.SOUERES, N.MANSARD, F.LAMIRAUX, J.P.LAUMOND, L.MARCHIONNI, H.TOME, F.FERRO

GEPETTO, Pal Robotics

Rapport LAAS N°17055, Mars 2017, 8p.

Lien : https://hal.archives-ouvertes.fr/hal-01485519

Diffusable

Plus d'informations

Abstract

The upcoming generation of humanoid robots will have to be equipped with state-of-the-art technical features along with high industrial quality, but they should also offer the prospect of effective physical human interaction. In this paper we introduce a new humanoid robot capable of interacting with a human environment and targeting a whole range of industrial applications. This robot is able to handle weights of 6 Kg with an outstretched arm, and has powerful motors to carry out movements unavailable in previous generations of humanoid robots. Its kinematics has been specially designed for screwing and drilling motions. In order to make interaction possible with human operators, this robot is equipped with torque sensors to measure joint effort and high resolution encoders to measure both motor and joint positions. The humanoid robotics field has reached a stage where robustness and repeatibility is the next watershed. We believe that, this robot has the potential to become a powerful tool for the research community to successfully navigate this turning point, as the humanoid robot HRP-2 was in its own time.

139336
17054
22/03/2017

A Kinodynamic steering-method for legged multi-contact locomotion

P.FERNBACH, S.TONNEAU, A.DEL PRETE, M.TAIX

GEPETTO

Rapport LAAS N°17054, Mars 2017, 7p.

Lien : https://hal.laas.fr/hal-01486933

Diffusable

Plus d'informations

Abstract

We present a novel method for synthesizing collision-free, dynamic locomotion behaviors for legged robots, including jumping, going down a very steep slope, or recovering from a push using the arms of the robot. The approach is automatic and generic: non-gaited motions, comprising arbitrary contact postures can be generated along any environment. At the core of our framework is a new steering method that generates trajectories connecting two states of the robot. These trajectories account for the state-dependent, centroidal dynamic constraints inherent to legged robots. The method, of low dimension, formulated as a Linear Program, is really efficient to compute, and can find an application in various problems related to legged locomotion. By incorporating this steering method into an existing sampling-based contact planner, we propose the first kinodynamic contact planner for legged robots.

139335
17049
21/03/2017

Optimal design of compliant walkers

G.BUONDONNO, J.CARPENTIER, G.SAUREL, N.MANSARD, A.DE LUCA, J.P.LAUMOND

Rome, GEPETTO, Univ of Cambridge

Rapport LAAS N°17049, Mars 2017, 7p.

Lien : https://hal.archives-ouvertes.fr/hal-01483567

Diffusable

Plus d'informations

Abstract

We present an optimization framework for the design and analysis of underactuated biped walkers, characterized by passive or actuated joints with rigid or non-negligible elastic actuation/transmission elements. The framework is based on optimal control, dealing with geometric constraints and various dynamic objective functions, as well as boundary conditions, which helps in selecting optimal values both for the actuation and the transmission parameters. Solutions of the formulated problems are shown for different kinds of bipedal architectures, and comparisons drawn between traditional rigid robots and compliant ones show the energy-efficiency of compliant actuators in the context of locomotion.

139321
17045
16/03/2017

Quelle place pour les robots humanoïdes dans nos maisons ? Une perspective épistémologique et sociale.

B.TONDU

GEPETTO

Rapport LAAS N°17045, Mars 2017, 17p.

Lien : https://hal.archives-ouvertes.fr/hal-01481029

Diffusable

Plus d'informations

Résumé

La robotique humanoïde peut-être considérée comme le but ultime d’une robotisation dont le robot industriel ne serait qu’une étape. De même que l’on parle d’une I.A. « faible » et d’une I.A. « forte », nous proposons d’opposer une robotique faible qui serait celle des robots industriels et une robotique forte qui serait celle des robots humanoïdes dont la capacité gestuelle se confondrait avec la capacité gestuelle humaine. Notre univers domestique est fait de machines conçues pour remplacer le travail du geste humain par le travail d’une machine non anthropomorphe, dont l’archétype est la machine à laver. Le robot domestique devra donc apprendre le maniement de ces machines domestiques qui, à la différence de la tâche qu’elles sont censées mimer, exigent peu de puissance mais une grande diversité gestuelle. Comment penser le devenir de la robotique humanoïde de service face à cette exigence d’habileté manuelle ? Il est bien difficile d’y répondre comme il était, à l’époque, bien difficile de cerner l’évolution de l’I.A. Ce n’est, selon nous, que par tâtonnements successifs que la robotique humanoïde prouvera sa valeur, entre relative spécialisation et relative standardisation de son environnement domestique où elle est censée s’intégrer. Enfin, il est encore difficile de dire si, dans un premier temps au moins, la robotique humanoïde domestique sera plus ostentatoire que réellement utile.

139284
17053
06/03/2017

Has uncanny valley feeling an impact on learning: A preliminary experiment performed in a french elementary school

B.TONDU

GEPETTO

Manifestation avec acte : R4L HRI Workshop on Robots for Learning ( ) 2017 du 06 mars au 06 mars 2017, Vienne (Autriche), Mars 2017, 10p. , N° 17053

Lien : https://hal.archives-ouvertes.fr/hal-01480569

Diffusable

Plus d'informations

Abstract

In order to test the eventual impact of uncanny feeling, as theorized by the Mori " s concept of uncanny valley, on learning, we proposed to a class of 26 pupils of mean age 8 years5 months to visualize The Polar Express and Dark Crystal movies and to test about four months after their ability to remember movie " s hero, story and end. The high degree of uncanny feeling initially produced by The Polar Express was verified by means of a questionnaire and discussion in the class performed just after the first scene of the movie was stopped. When some weeks after the entire movies was shown to the pupils, this uncanny feeling appears to be weaker and when finally pupils were tested on their memories concerning both movies, answers got in The Polar Express case did not appear to be less relevant than those got for Dark Crystal. The reported experiment suggests that highly human-like faces do not prevent to transmit information to be learned by pupils, which tends to indicate that faces of future teaching robots could take human form.

139333
17018
10/02/2017

Formation Doctorale "Éthique de la recherche et Intégrité scientifique": Une formation proposée par l'école des docteurs de Toulouse

C.TESSIER, F.VIALLET, A.MARY, B.TONDU, L.BAUDOUIN, P.DUCOURNAU, I.QUEINNEC, G.MOCHEL, C.VARENNE, N.DESSENS

ONERA, EFTS, LIRCES, Nice, GEPETTO, MAC, LISST, Toulouse, URFIST-UFTMIP, TRACES, UT2J

Rapport LAAS N°17018, Février 2017, 56p.

Lien : https://hal.archives-ouvertes.fr/cel-01452867

Diffusable

Plus d'informations

Résumé

Les écoles doctorales du site toulousain ont demandé fin 2015 que l’École Des Docteurs pilote un groupe de travail permettant de développer une formation à « l’Éthique de la Recherche et à l’intégrité scientifique ». Un groupe de chercheurs et d’enseignants-chercheurs, identifiés grâce à un appel à participation diffusé dans les écoles doctorales toulousaines, a développé un « kit de formation » comprenant : un diaporama support de cours, des propositions de dilemmes éthiques formatés et ce guide pratique à destination des formateurs. Ces documents sont inspirés du Cahier « Proposition de formation doctorale ; Initiation à l’éthique de la recherche scientifique » de la CERNA (Commission de réflexion sur l’Éthique de la Recherche en sciences et technologies du Numérique d’Allistene), des propositions des différents acteurs du groupe de travail, issus de disciplines diverses et du retour d’expérience des 8 premières journées de formation données à des groupes de doctorants issus de différentes écoles doctorales . L’Université de Toulouse est la première en France où cette formation est dispensée de manière transversale à tous les doctorants. Les échanges interdisciplinaires (entre formateurs et entre étudiants) sont particulièrement fructueux pour la dynamique de formation, qui se doit d’être participative. Cette formation est donnée en binôme, chaque binôme comprenant un intervenant issu des disciplines science, technologie, santé, ingénierie et un intervenant issu des sciences humaines et sociale, droit économie, gestion. Pendant l’année universitaire 2015-2016, cette formation a été testée lors de 4 sessions. Elle a été reconduite à plus grande échelle pour l’année universitaire 2016-2017, avec des binômes « ancien formateur/nouveau formateur », permettant ainsi d’assurer la formation des formateurs tout en conservant la cruciale complémentarité disciplinaire. À ce jour sur le site toulousain, ce sont donc une douzaine d’intervenants chercheurs, enseignants-chercheurs, ingénieurs et personnels techniques de recherche et de formation qui ont formé plus de 130 doctorants de disciplines et horizons différents.

138962
17024
01/02/2017

COCoMoPL: A Novel Approach for Humanoid Walking Generation Combining Optimal Control, Movement Primitives and Learning and its Transfer to the Real Robot HRP-2

D.CLEVER, M.HARANT, K.MOMBAUR, M.NAVEAU, O.STASSE, D.ENDRES

IWR, Heidelberg, GEPETTO, Marbourg

Revue Scientifique : IEEE Robotics and Automation Letters, Vol.2, N°2, pp.977-984, Février 2017 , N° 17024

Lien : https://hal.archives-ouvertes.fr/hal-01459840

Diffusable

Plus d'informations

Abstract

COCoMoPL [6] is a recently developed approach Combining Optimal Control, Movement Primitives and Learning for the generation of humanoid walking motions. It solves optimal control problems based on detailed dynamic models of the robot for a variety of walking parameters and uses the solutions as training data to create movement primitives that are very close to feasibility and optimality. These can be employed to synthesize complex walking sequences for humanoid robots online in a very efficient way. We demonstrate, for the first time, that COCoMoPL works on a real humanoid robot, here HRP-2 with 36 DOF and 30 position controlled actuators. To this end, it was necessary to significantly extend the existing approach by including transition steps into the training data, modify the movement primitives (MP) to admit these transitions, improve the representation of the ZMP MPs and tighten the transition conditions at the beginning and end of steps. We present a thorough validation of the method in simulation and on the real robot for a challenging sequence of movements. We also compare the characteristics of movements after each step of the methodology.

139062
16496
27/01/2017

Continuous legged locomotion planning

N.PERRIN, C.OTT, J.ENGLSBERGER, O.STASSE, F.LAMIRAUX, D.G.CALDWELL

ISIR, Jussieu, DLR, GEPETTO, IIT, Genova

Rapport LAAS N°16496, doi 10.1109/TRO.2016.2623329, Janvier 2017, 7p.

Lien : http://hal.upmc.fr/hal-01433866

Diffusable

Plus d'informations

Abstract

While only continuous motions are possible, the way contacts appear and disappear confers to legged locomo-tion a characteristic discontinuous nature that is traditionally shared by the algorithms used for legged locomotion planning. In this paper we show that this discontinuous nature can disappear if the notion of collision is well redefined, and we efficiently solve two different practical problems of legged locomotion planning with algorithms based on an approach that establishes a bridge between discrete and continuous planning. The first problem consists of reactive footstep planning with a biped robot, and the second one of non-gaited locomotion planning with a hexapod.

138684
17002
26/01/2017

Spray-coated carbon nanotube carpets for creeping reduction of conducting polymer based artificial muscles

A.SIMAITE, A.DELAGARDE, B.TONDU, P.SOUERES, E.FLAHAUT, C.BERGAUD

MEMS, ELIA, GEPETTO, CIRIMAT

Revue Scientifique : Nanotechnology, Vol.28, N°2, 025502p., Janvier 2017 , N° 17002

Lien : https://hal.laas.fr/hal-01413022

Diffusable

Plus d'informations

Abstract

It is often observed that during cyclic actuation conducting polymer based artificial muscles are continuously creeping from the initial movement range. One of the likely reasons of such behaviour is unbalanced charging during conducting polymer oxidation and reduction. In order to improve the actuation reversibility and subsequently the long time performance of ionic actuators, we suggest to use spray-coated carbon nanotube (CNT) carpets on the surface of the conducting polymer electrodes. We show that carbon nanotubes facilitate conducting polymer redox reaction and improve its reversibility. Consequently, in the long term, charge accumulation in the polymer film is avoided leading to significantly improved long term performance during cycling actuation.

138654
Les informations recueillies font l’objet d’un traitement informatique destiné à des statistiques d'utilisation du formulaire de recherche dans la base de données des publications scientifiques. Les destinataires des données sont : le service de documentation du LAAS.Conformément à la loi « informatique et libertés » du 6 janvier 1978 modifiée en 2004, vous bénéficiez d’un droit d’accès et de rectification aux informations qui vous concernent, que vous pouvez exercer en vous adressant à
Pour recevoir une copie des documents, contacter doc@laas.fr en mentionnant le n° de rapport LAAS et votre adresse postale. Signalez tout problème de dysfonctionnement à sysadmin@laas.fr. http://www.laas.fr/pulman/pulman-isens/web/app.php/