[PDF] The Cog Project: Building a Humanoid Robot





Previous PDF Next PDF



Quelques dates clés Quelques définitions

1923 Invention du mot « robot » (pièce de théâtre de Karel Capek) 1993 Cog Humanoid Robotics Group



Projet Robotique Lévolution des robots 3ème

A l'aide des sites Internet ressources et des dates clés de l'histoire des robots 1993 Cog



Automatisme et Robotique RESSOURCE

1993 Cog Humanoid Robotics Group



The Cog Project: Building a Humanoid Robot

MIT Artificial Intelligence Lab http://www.ai.mit.edu/projects/cog/. Abstract. ... group began the construction of a humanoid robot.



_Histoire des robots

1993 Cog Humanoid Robotics Group



Evolution de Evolution de lobjet technique Travail 6 technique 3

A l'aide des dates clés de l'histoire des robots 1993 Cog Humanoid Robotics Group



Poppy: open-source 3D printed and fully-modular robotic platform

13 juil. 2015 à l'Humanoid Robotics Group de l'Institut de Technologie du Massachusetts(MIT). ... L'approche scientifique des robots Cog est orientée vers ...







Evolution de lobjet technique Travail 1 3

A l'aide des dates clés de l'histoire des robots (doc 3/3) compléter la frise 1993 Cog



TheCogProject:BuildingaHumanoidRobot

Research in humanoid robotics has uncovered a variety of new problems andafewsolutionstoclassicalproblemsinroboticsarti?cialintelligenceand control theory This research draws upon work in developmental psychology ethologysystemstheoryphilosophyandlinguisticsand throughthe process



The Cog Project: Building a Humanoid Robot - MIT CSAIL

group began the construction of a humanoid robot This research project has two goals: an engineering goal of building a prototype general purpose ?exible and dextrous autonomous robot and a scienti c goal of understanding human cognition (Brooks & Stein 1994) Recently many other research groups have begun to construct integrated hu-



1 Introduction - Massachusetts Institute of Technology

In the Humanoid Robotics Group at MIT CSAIL we are currently developing a new force sensing and compliant humanoid namedDomo Domois to be a research platform for exploring issues in general dexterous manipulation visual perception and learning This project is currently in the design and development phase

What is a Cog robot?

Cog was a humanoid robot designed by Rodney Brooks's group at MIT as a platform to study robot cognition. It could track faces, grasp objects, and, perhaps most famously, play with a Slinky. How do you like this robot? Would you want to have this robot? Did You Know? Able to recognize objects and reach for a visual target.

What is a humanoid robot?

In the summer of 1993, ourgroup began the construction of a humanoid robot. This research project hastwo goals: an engineering goal of building a prototype general purpose ?exibleand dextrous autonomous robot and a scientifc goal of understanding humancognition (Brooks & Stein 1994).

What was the first humanoid robot to use Series Elastic actuators?

Cog was one of the first humanoid robots to use series elastic actuators: The motors on the arms were connected to the joints in series with a torsional spring, which protected the gearbox and provided compliance and more safety for people interacting with the arms.

What is an upper-torso humanoid robot called?

Abstract. To explore issues of developmental structure, physical em-bodiment, integration of multiple sensory and motor systems, and socialinteraction, we have constructed an upper-torso humanoid robot calledCCog.

TheCogProject:BuildingaHumanoidRobot

BrianScassellati,MatthewM.Williamson

MITArticialIntelligenceLab

545TechnologySquare

CambridgeMA02139,USA

http://www.ai.mit.edu/projects/cog/ re tem.

1Introduction

exible cognition(Brooks&Stein1994). uencedourresearch theopenproblemsthathaveyettobeaddressed.

2Methodology

andmultimodalintegration. thenbrie constructinghumanoidroboticsystems. sciencerefutestheseassumptions. task. centralmodelofvisualspace. science. ratherthanasinglemonolithicone. sameproblem.

2.2EssencesofHumanIntelligence

ofcreatinghuman-likeintelligence. competencyofthesystem. uxofstimula- mentalprogression.

1998).

taskconstraints.

1998a,Williamson1998b).

onemodalitycananddoin usingvestibularfeedback. integration. thedevelopmentofearlyAI.

3Hardware

3.1ComputationalSystem

3.2PerceptualSystems

digitalsignalprocessornetwork. processornetwork. robot.

3.3MotorSystems

1986).

ectsoutoftheway.Thedis- 31"
degrees)andbodytwist(120degrees)

3.4DevelopmentPlatforms

designcanbefoundinScassellati(1998a). ofthebehaviorengine. rentemotionalstateoftherobot.

4CurrentLong-TermProjects

4.1JointAttentionandTheoryofMind

state. approaches. thatarecriticaltoinfantdevelopment. itsactionsin guidethelearningprocess. appropriateemotiveandexpressivecues. interactionthattherobotrequires. itsemotiveactsin fortherobottolearnhowitsemotiveactsin uencethebehaviorofthecare- satisfyitsdrives.

4.3DynamicHuman-likeArmMotion

thenaturaldynamicsoftherobottoobtain exibleandrobustmotionwithout complexcomputation.

4.4Multi-modalCoordination

eachother. example). isbynomeanstrivial. limits,andactuatoraccuracy. ofredundantinformationproducedbycon uentdatastreams.Anycorrelation betweeneventswithtimedelays. lookingtasks.Postureisnotmerelyare exivecontrol;ithasfeed-forwardcom- amountofmulti-modalintegration.

5CurrentTasks

y http://www.ai.mit.edu/projects/cog/.

5.1Visual-motorRoutines

ments(thevestibulo-ocularre theseeyemotions. imageprojectionchangesslowly. smoothpursuit.

Vestibular-ocularandOpto-kineticRe

exes:Thevestibulo-ocularre- headmoves.Thevestibulo-ocularre ex(VOR)stabilizestheeyesduringrapid bymeasuringtheoptic otolithorgans). ofoptic optic owestimateisa necktothattarget.

5.2Eye/NeckOrientation

someconstantk.2 neckisinmotion:thevestibulo-ocularre exoraneerencecopysignalofthe additionsnecessary.Becausethere exusesgyroscopefeedbacktomaintainthe k,wherekisthesameconstant ex,the

INPUTTONIC

c INPUT TONIC c v1 1 2 v2bb hj [gj]- wy2wwy1hj [gj]+ y1 y2 youtPROPRIOCEPTIVE

INPUTgjOUTPUT

g y

5.3DynamicOscillatorMotorControl

bedescribedby: u i=ki(vii)bi_i(1) inunstructuredenvironments. neuronisgovernedbythefollowingequations:

1_x1=x1v1![x2]+j=n

j=1hj[gj]++c(2)

2_v1=v1+[x1]+(3)

1_x2=x2v2![x1]+j=n

j=1hj[gj]+c(4)

2_v2=v2+[x2]+(5)

y i=[xi]+=max(xi;0)(6) y out=y1y2(7)

Matsuoka(1987).

Oneneuron

theoverallarmmotion. synchronywiththeshoulder.

01234567-20

-10 0 10 20 30

Time seconds

joint angles

With force feedback

01234567-20

-10 0 10 20 30

Time seconds

joint angles

Without force feedback

sh angle el angle speed onlywiththeproprioceptivefeedback. coordinationisshowninFigure8. oftheoscillations.

5.4PointingtoaVisualTarget

024681012-40

-20 0 20 40
60
time - seconds equilbrium angle

024681012-40

-20 0 20 40
60
time - seconds equilibrium angle left arm right arm feedback gain appropriatefortrainingtheballisticmap. ail- throughitsworkspace. simplieslearning. ofactiveresearch. containingtheeyeforfurtherprocessing.

5.6Imitatingheadnods

mustcomefromaface-likeobject. andfacialexpressions.Thesesystemsin uenceeachothertoestablishandmain- stimulithatcanbein robot'sfacialexpressionsre driveswithinhomeostaticranges. occurs,therobot'sexpressionre

0204060801001201401601802000

500
1000
1500
2000

2500Interaction with face

Time (seconds)

Activation Level

Anger

Disgust

Interest

Sadness

Happiness

020406080100120140160180200

-2000 -1000 0 1000
2000

Time (seconds)

Activation LevelSocial driveSocialize behaviorFace stimulus toanger. toomuchandoverstimulatingtherobot. interactionin

6FutureResearchDirections

4.4.

6.1Coherence

awaywhichhasproducedanyvalidsolutions.

6.2OtherPerceptualSystems

taste.

6.3DeeperVisualPerception

ex,andver- environment. neitherisnecessarilythecorrectapproach. pathtakenbychildren. variedlighting.4

6.4ASenseofTime

the owoftimeweashumanbeingsexperience. programanddata,andtheprogramhasanatural owoftimethatitcanthen theproblemofhowepisodicmemorymightarise.

7Acknowledgments

Grant(No.N00014-95-1-0600).

References

manipulator,MITPress,Cambridge,MA.

JournalofCognitiveNeurosciencepp.66{80.

pp.668{673. pp.720{726. pp.727{733.

SimulatedAdaptiveBehaviorWorkshop'.

(AAAI-98)'.

1(1),7{25.

Press.

AcademyofSciences,pp.637{676.

Cod,Massachusetts,pp.94{103.

FallSymposium',AAAIPress.

York.

3rdedn,AppletonandLange.

14(4),260{286.

guage,Elsevier,NewYork,NY.

SceneAnalysis(CASA'97)',IJCAI-97.

Automation(ICRA-98)',IEEEPress.

Press,pp.191{206.

modelbasedonthevestibulo-ocularre ex',Nature260,159{161. (AAAI-97)',pp.748{754.

Intelligence(AAAI-97)',pp.454{459.

465.

RoboticsandAutomation'.

intelligenceatMIT,1970-71'. 2743.
exes:Integrationof

Laboratory.

ConfonRoboticsandAutomation'.

AAAIPress.

MITArticialIntelligenceLab.Insubmission.

AAAIPress.

tionalConferenceonAutonomousAgents'.

Automation',Vol.3,pp.913{918.

22(8{9),805{817.

sachusettsInstituteofTechnology. imentalPsychology18,643{62.

UniversityPress,pp.321{348.

Press,pp.349{370.

Oxford.

quotesdbs_dbs44.pdfusesText_44
[PDF] les rovers spirit et opportunity sur mars nasa

[PDF] hume la regle du gout

[PDF] marché de la robotique domestique

[PDF] robot industriel

[PDF] wowwee rovio

[PDF] wowwee rovio 2

[PDF] robomow

[PDF] rovio wowwee software

[PDF] humminbird helix 5 francais

[PDF] humminbird helix 5 mode d'emploi

[PDF] humminbird helix 5 manuel

[PDF] humminbird helix 5 di gps manual

[PDF] humminbird helix 5 di gps manuel

[PDF] programme pour robot suiveur de ligne

[PDF] reglage sondeur humminbird helix 5