Embodied Intelligence: The four messages
3 likes2,234 views
The document discusses the concept of embodied intelligence, emphasizing the importance of physical embodiment in robotics and intelligence development. It outlines four key messages related to the relationship between brain function, body morphology, and environmental interaction, along with future directions towards 'soft robotics.' Various case studies and applications illustrate how understanding embodiment can lead to advancements in robot design and functionality.
![Sensory-motor
coordination
“We begin not with a sensory stimulus, but with a
(“active perception”)
sensory-motor coordination […] In a certain sense it is
the movement which is primary, and the sensation
which is secondary, the movement of the body, head,
and eye muscles determining the quality of what is
experienced. In other words, the real beginning is with
the act ofthe stimulations which thenot a sensation of
“Since all seeing; it is looking, and organism receives
light.” (“Thebeen possiblepsychology,” John Dewey,
have in turn reflex arc in only by its preceding
1896)
movements which have culminated in exposing the
receptor organ to external influences, one could also
say that behavior is the first cause of all the
stimulations.” (“The structure of Behavior,” Maurice 64
Merleau-Ponty, 1963)](https://image.slidesharecdn.com/swissnexsfoslides2011-120202181725-phpapp01/85/Embodied-Intelligence-The-four-messages-78-320.jpg)
Embodied Intelligence: The four messages
- 1. Embodied intelligence: The four messages With a case study on ECCERobot Swissnex San Francisco ... 19 January 2012 Rolf Pfeifer, NCCR National Competence Center Robotics, Switzerland
- 2. Thanks to ... Hajime Asama Robert Full Lukas Rudolf Bannasch Gabriel Gomez Lichtensteiger Josh Bongard Fumio Hara Hod Lipson Simon Bovet Alejandro Max Lungarella Rodney Brooks Hernandez Ren Luo Weidong Chen Owen Holland Barbara Mazzolai Steve Collins Koh Hosoda Shuhei Miyashita Holk Cruse Fumiya Iida Toshi Nakagaki Paolo Dario Auke Ijspeert Norman Packard Raja Dravid Takashi Ikegami Mike Rodney Douglas Masayuki Inaba Rinderknecht Peter Akio Ishiguro Roy Ritzmann Eggenberger Oussama Kathib Andy Ruina Andreas Engel Alois Knoll Giulio Sandini Martin Fischer Maarja Kruusma Olaf Sporns Dario Floreano Yasuo Kuniyoshi Luc Steels Toshio Fukuda Cecilia Laschi Kasper Stoy
- 3. for their ideas Hajime Asama Robert Full Lukas Rudolf Bannasch Gabriel Gomez Lichtensteiger Josh Bongard Fumio Hara Hod Lipson Simon Bovet Alejandro Max Lungarella Rodney Brooks Hernandez Ren Luo Weidong Chen Owen Holland Barbara Mazzolai Steve Collins Koh Hosoda Shuhei Miyashita Holk Cruse Fumiya Iida Toshi Nakagaki Paolo Dario Auke Ijspeert Norman Packard Raja Dravid Takashi Ikegami Mike Rodney Douglas Masayuki Inaba Rinderknecht Peter Akio Ishiguro Roy Ritzmann Eggenberger Oussama Kathib Andy Ruina Andreas Engel Alois Knoll Giulio Sandini Martin Fischer Maarja Kruusma Olaf Sporns Dario Floreano Yasuo Kuniyoshi Luc Steels Toshio Fukuda Cecilia Laschi Kasper Stoy
- 4. for their ideas Hajime Asama Robert Full Lukas Rudolf Bannasch Gabriel Gomez Lichtensteiger Josh Bongard Fumio Hara Hod Lipson Simon Bovet Alejandro Max Lungarella Rodney Brooks Hernandez Ren Luo Weidong Chen Owen Holland Barbara Mazzolai Steve Collins Koh Hosoda Shuhei Miyashita Holk Cruse Fumiya Iida Toshi Nakagaki Paolo Dario Auke Ijspeert Norman Packard Raja Dravid Takashi Ikegami Mike Rodney Douglas Masayuki Inaba Rinderknecht Peter Akio Ishiguro Roy Ritzmann Eggenberger Oussama Kathib Andy Ruina Andreas Engel Alois Knoll Giulio Sandini Martin Fischer Maarja Kruusma Olaf Sporns Dario Floreano Yasuo Kuniyoshi Luc Steels Toshio Fukuda Cecilia Laschi Kasper Stoy
- 5. Goals buzzword “embodiment” seeing things differently “mind set” for design 5
- 6. Contents • Design principles and “the four messages of embodiment” • A bit of background • Where are we going? — “Soft robotics” • Take home message 6
- 7. Getting into the spirit of embodiment the role of the brain in understanding behavior? 7
- 10. The spirit of embodiment 8
- 11. The spirit of embodiment 8
- 12. The spirit of embodiment 8
- 13. “Crazy Bird” — Morphology, Control loosely hanging feet rubber/plastic Design and construction: Mike Rinderknecht 9
- 14. “Crazy Bird” — Morphology, Control loosely hanging feet rubber/plastic Design and construction: behavior of “Crazy Mike Rinderknecht Bird”: emergent 10
- 15. Message 1: Physical embedding Studying brain (or control) not sufficient: Understanding of • embedding of brain into organism • organism’s morphological and material properties • interaction with environment required 11
- 16. Let me be clear The brain is important! 12
- 17. Let me be clear The brain is important! but not the whole story ... 13
- 18. Contents • Design principles and “the four messages of embodiment” • A bit of background • Where are we going? — “Soft robotics” • Take home message 14
- 19. Artificial Intelligence — goals 1.Understanding 2.Making biological abstractions, systems developing theory humans animals 3.Applications beer-serving robot 15 vacuum cleaner
- 20. The synthetic methodology Slogan: “Understanding by building” modeling behavior of interest abstraction of principles robots as tools for scientific investigation 16
- 21. The synthetic methodology New scientific paradigm (general trend) beyond classical, analytical sciences novel types of experiments 17
- 22. Zurich AI Lab robots Ms. Gloria Rufus T. Teasdale Firefly Didabot Famez Sita Morpho
- 23. Zurich AI Lab robots Amouse SahabotI/II Melissa Tripp Samurai Analogrob Dexterolator Stumpy Eyebot Mindstorms Kheperas Mitsubishi Forkleg
- 24. Zurich AI Lab robots Stumpy, Monkey, Puppy, Min-dog, Wheeled Walker, Mini-Stumpy, Wanda, Dumbo, Rabbit
- 25. Zurich AI Lab robots 21
- 26. AI Lab Robots (exploration of 22
- 27. Zurich AI Lab Robots (EU-Locomorph) 23
- 28. Two views of intelligence classical: cognition as computation embodiment: cognition as emergent from movement, locomotion, manipulation Illustrations 24 by Shun Iwasawa
- 29. Classical approach view: thinking/brain: centralized control model: input - processing - output how else could it be? 25
- 30. Successes and failures of the classical successes failures applications foundations of (e.g. Google) behavior chess natural forms of intelligence consumer electronics interaction with real world manufacturing 26
- 31. Successes and failures of the classical successes failures applications foundations of (e.g. Google) behavior chess natural forms of intelligence consumer electronics interaction with real world factory 27
- 32. Where is the problem? • inappropriate view of intelligence as: “input-processing-output”(computer metaphor) • neglect of interaction with real world what to do? 28
- 33. Two views of intelligence classical: cognition as computation embodiment: cognition as emergent from movement, locomotion, manipulation Illustrations 29 by Shun Iwasawa
- 34. Relation to thinking/intelligence? “Why do plants not have brains?” 30
- 35. Relation to thinking/intelligence? “Why do plants not have brains? The answer is actually quite simple: they don’t have to move.” Lewis Wolpert, UK —> evolutionary perspective 31
- 36. Back to the “four messages of 32
- 37. Message 2: Real/Artificial worlds Understanding the differences between artificial/constructed (e.g. industrial) worlds and real worlds (e.g. downtown area, school, home) —> different requirements for robots 33
- 38. industrial robots natural (“hard”) systems (“soft”) human s industrial robots 34
- 39. industrial natural robots systems principles: (“soft”) - high predictability - strong, fast, precise motors - centralized control - optimization industrial robots 35
- 40. industrial natural robots systems principles: - inprecise (“soft”) human - compliant s - reactive - coping with uncertainty industrial robots 36
- 41. industrial natural robots systems principles: - inprecise (“soft”) human - compliant s - reactive - coping with uncertainty no direct transfer of methods 37
- 42. Transfer of methods? Sony Qrio: high stiffness centralized control conputationally intensive 38
- 43. Transfer of methods? Sony Qrio: high stiffness centralized control conputationally intensive 38
- 44. Transfer of methods? Sony Qrio: high stiffness centralized control conputationally intensive 38
- 45. Transfer of methods? Sony Qrio: high stiffness centralized control conputationally intensive 38
- 46. By comparison: The “Passive Dynamic Design and construction: Ruina, Wisse, Collins: Cornell University Ithaca, New York The “brainless” robot”: walking without control 39
- 47. By comparison: The “Passive Dynamic Design and construction: Ruina, Wisse, Collins: Cornell University Ithaca, New York The “brainless” robot”: walking without control 39
- 48. By comparison: The “Passive Dynamic self-stabilization Design and construction: Ruina, Wisse, Collins: Cornell University Ithaca, New York The “brainless” robot”: walking without control 40
- 49. By comparison: The “Passive Dynamic self-stabilization Design and construction: Ruina, Wisse, Collins: Cornell University Ithaca, New York The “brainless” robot”: walking without control 40
- 50. Overall scheme: Self- stabilization in the Passive Pfeifer et al.,Science, 16 Nov. 2007 41
- 51. Overall scheme: Self- stabilization in the Passive self-stabilization Pfeifer et al.,Science, 16 Nov. 2007 42
- 52. Short question memory for walking? 43
- 53. Contrast: Full control — “hard” Honda Asimo Sony Qrio 44
- 54. Extending the ecological niche adding a reflex Design and construction: Martijn Wisse, Delft University ompare: human walking self-stabilization 45
- 55. Extending the ecological niche adding a reflex Design and construction: Martijn Wisse, Delft University ompare: human walking self-stabilization 45
- 56. Implications of embodiment “Denise” self-stabilization Pfeifer et al.,Science, 16 Nov. 2007 46
- 57. Message 3: Task distribution Task distribution between brain (control), body (morphology, materials), and environment Principle of ecological balanceof Principle cheap design 47
- 58. Message 3: Task distribution Task distribution between brain (control), body (morphology, materials), and environment morphological computation 48
- 59. Message 3: Task distribution Task distribution between brain (control), body (morphology, materials), and environment no clear separation between control and hardware (“soft robotics”) rethink classical control 49
- 60. The “robot frog” driven by pneumatic actuators Design and construction: Ryuma Niiyama and Yasuo Kuniyoshi University of Tokyo 50
- 61. The “robot frog” driven by pneumatic actuators Design and construction: Ryuma Niiyama and Yasuo Kuniyoshi University of Tokyo 50
- 62. “Stumpy”: task distribution almost brainless: 2 actuated joints springy materials surface properties of feet Design and construction: Raja Dravid, Chandana Paul, Fumiya Iida self-stabilization 51
- 63. The dancing robot “Stumpy” Collaboration with Louis-Philippe Demers, Nanyang Technological University, Singapore Movie: Dynamic Devices and 52 AILab, Zurich
- 64. The dancing robot “Stumpy” Collaboration with Louis-Philippe Demers, Nanyang Technological University, Singapore Movie: Dynamic Devices and 52 AILab, Zurich
- 65. Outsourcing functionality Mini-rHex Design and construction: Robin Guldener, Lijin Aryananda soft, flexible, elastic materials 53
- 66. Outsourcing functionality Mini-rHex Design and construction: Robin Guldener, Lijin Aryananda soft, flexible, elastic materials 53
- 67. Specifically: Orchestration of • stably grasping hard object • other manipulation tasks • morphological computation: exploiting morphology/ materials for control Design and construction: Koh Hosoda 54
- 68. Exploiting morphology: managing complex pictures and ideas: courtesy Roy Ritzmann Case Western Reserve University 55
- 69. Exploiting morphology: managing complex pictures and ideas: courtesy Roy Ritzmann Case Western Reserve University 55
- 70. “Outsourcing” functionality: • brain: 1 Million neurons (rough estimate) • descending neurons: 200 (!) • brain: - cooperation with local circuits - morphological changes (shoulder joint) Watson, Ritzmann, Zill & Pollack, 2002, J Comp Physiol A 56
- 71. Effects of morphology change descending brain neurons: 200 (!) 1 Million neurons 57
- 72. Climbing over obstacles • CPG on flat ground • get hight estimate from antenna • change configuration of shoulder joint —> morphological computation • CPG continue to function as before (don’t “know” about climbing) • brain-body cooperation 58
- 73. Towards “cognitive” robots 59
- 74. Adding sensors: generation of sensory stimulation through action • knowledge about environment: pressure, haptic, acceleration, vision, ... • knowledge about own body: angle, torque, force, vestibular, … the super-compliant “soft” robot ECCE
- 75. Message 4: Physical dynamics and Induction of patterns of sensory stimulation through physical interaction with environment raw material for information processing of brain (control) induction of correlations (information structure) through sensory-motor 61
- 76. Message 4: Physical dynamics and Induction of patterns of sensory stimulation through physical interaction with environment raw material for information processing of brain (control) induction of correlations (information Principle of information structure) through sensory-motor 62 self-structuring
- 77. Essence • self-structuring of sensory data through — physical — interaction with world • physical process — not computational pre-requisite for learning Inspiration: John Dewey, 1896 (!) Merleau-Ponty, 1963 63 Bajcsy, 1963; Aloimonos, 1990; Ballard, 1991 Sporns, Edelman, and co-workers
- 78. Sensory-motor coordination “We begin not with a sensory stimulus, but with a (“active perception”) sensory-motor coordination […] In a certain sense it is the movement which is primary, and the sensation which is secondary, the movement of the body, head, and eye muscles determining the quality of what is experienced. In other words, the real beginning is with the act ofthe stimulations which thenot a sensation of “Since all seeing; it is looking, and organism receives light.” (“Thebeen possiblepsychology,” John Dewey, have in turn reflex arc in only by its preceding 1896) movements which have culminated in exposing the receptor organ to external influences, one could also say that behavior is the first cause of all the stimulations.” (“The structure of Behavior,” Maurice 64 Merleau-Ponty, 1963)
- 79. Contents • “The four messages of embodiment” • A bit of background • Where are we going? — “Soft robotics” • Take home message 65
- 80. “Soft robotics” Hypothesis: The next generation of robots will be of the “soft” kind. Advances in “soft technology” will lead to a quantum leap in intelligent robotics. Theoretical underpinnings: The key to “soft robotics” will be an understanding of embodiment. 66
- 81. Application areas • safe interaction with humans • next level factory automation • manipulation for assembly and surgery • therapy and human assistance • mobility over different terrain • companion robot • entertainment 67 • many others
- 82. “Soft Robotics” Soft to touchSoft movement oft interaction Emotions S 68
- 83. “Soft Robotics” Soft to touch Soft movement Soft interaction Emotions - materials - elastic - soft movements friendly - - soft skin compliant - social and interaction - deformable materials for cognitive skills with human tissue muscles and - reactive - facial - fur tendons - soft materials expression - variable compl. - body actuators posture - expl. passive dynamics 69
- 84. “Soft Robotics” Soft to touch Soft movement Soft interaction Emotions - materials - elastic - soft movements friendly - - soft skin compliant - social and interaction - deformable materials for cognitive skills with human tissue muscles and - reactive - facial - fur tendons - soft materials expression - variable compl. - body actuators posture - expl. passive dynamics 70
- 85. The super-compliant “soft” robot ECCE Design and construction: Rob Knight — robotstudio, Geneva Richard Newcombe — Imperial College ECCE — Embodied Cognition in a Compliantly Engineered Robot Anthropomor 71 phic design
- 86. The super-compliant “soft” robot ECCE ECCE — Embodied Cognition in a Compliantly Engineered Robot 72
- 87. Techfest 2011, IIT Bombay Embodied Intelligence Switzerland 73
- 88. i-Days, Lucern Switzerland 74
- 89. Hannover Fair, ICT Brussels, Science Fair St. 75
- 90. ECCE at Chinese Academy of Science, 76
- 91. ECCE in Singapore, 2011 UBS and Nanyang 77
- 92. ECCE with former president of Switzerland: Innovation Fair Design and construction: der “bionische Roboter” Rob Knight — robotstudio, Geneva Richard Newcombe — Imperial College Owen Holland — Essex/Sussex University ECCE — Embodied Cognition in a Compliantly Engineered Robot Anthropomor 78 phic design
- 93. The super-compliant “soft” robot ECCE Anthropomo rphic design 79
- 94. The super-compliant “soft” robot ECCE Anthropomo rphic design 79
- 95. “Soft robotics”: sample applications Receptionist at World Expo - Design and construction: Osaka University, and Kokoro Dreams Robot Teacher Saya - Design and construction: Hiroshi Kobayashi, Univ. of Science, Tokyo 80
- 96. “Soft robotics”: sample applications Receptionist at World Expo - Design and construction: Osaka University, and Kokoro Dreams Robot Teacher Saya - Design and construction: Hiroshi Kobayashi, Univ. of Science, Tokyo 80
- 97. Support Suits Exoskeletons paralyzed individual to climb Breithorn (Switzerland) 81 HAL, the “Hybrid Exoskeleton Assistive Limb ®” Cyberdyne Inc.
- 98. The “power: of materials: design and construction: Marc Ziegler, AI Lab, UZH materials! 82
- 99. The “power: of materials: design and construction: Marc Ziegler, AI Lab, UZH materials! 82
- 100. Exploiting materials: Octopus (EU Project) Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) Tao Li (UZH)
- 101. Octopus arm movements Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 84
- 102. Octopus arm movements Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 84
- 103. Grasping a bottle Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 85
- 104. Grasping a bottle Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 85
- 105. Pulling a string Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 86
- 106. Pulling a string Octopus Arm Design and construction: Matteo Cianchetti (SSSA) Cecilia Laschi (SSSA) control: Kohei Nakajima (AI 86
- 107. 87
- 108. The Jaeger/Lipson “coffee 88
- 109. The Jaeger/Lipson “coffee 89
- 110. The Jaeger/Lipson “coffee 89
- 111. Implications for automation? 90
- 112. Food handling • considerable variation —> unpredictability • extremely delicate • materials crucial 91
- 113. Recall: industrial robots (“hard”) vs. natural systems Prinzipien: - high predictability - strong, precise, fast motors - centralized control - optimization industrial —> reduced robots predictability 92
- 114. New manipulation skills 93
- 115. Outsourcing? Insourcing? Chinese textile workers discover their power 94
- 116. The next “industrial revolution” beyond traditional manufacturing: new manipulation skills new manufacturing hard robotics softbots technology new industrial revolution OCTOPUS arm prototype Rodney Brooks U-Tokyo 95 robot “frog” Festo Bionic ECCE Handling assistant the super-compliant robot
- 117. Contents • “The four messages of embodiment” • A bit of background • Where are we going? — “Soft robotics” • Take home message 96
- 118. Summary and conclusions Key to “soft robotics”: understanding of “embodiment” —> the “four messages” 97
- 119. “Soft robotics” Hypothesis: The next generation of robots will be of the “soft” kind. Advances in “soft technology” will lead to a quantum leap in intelligent robotics. Theoretical underpinnings: The key to “soft robotics” will be an understanding of embodiment. 98
- 120. “Soft robotics” • central role of materials! • new notion of control (morphological computation; “orchestration”) • no clear separation between controller and to-be-controlled 99
- 121. “Soft robotics” • central role of materials! • new notion of control (morphological computation; “orchestration”) • no clear separation between better robots controller and to-be-controlled better life 100
- 122. “Soft robotics” Hypothesis: The next generation of robots will be of the “soft” kind. Advances in “soft technology” will lead to a quantum leap in intelligent robotics. Theoretical underpinnings: The key to “soft robotics” will be an understanding of embodiment. 101
- 123. Summary: The four messages of Message 1: Physical embedding Understanding brain not enough; morphology materials; embedding Message 2: Real/Artificial worlds Fundamental differences industrial and real- world environments Message 3: Task distribution Cooperation - brain, body, environment 102 Message 4: Physical dynamics and information structure Induction of information structure;
- 124. Like to know more? 103
- 125. Visit us in Zurich 104
- 126. 105
- 127. Locomor Research program ph Morpho-function Octopus artificial dynamic movement Scalable morphogenesis Locognit NCCR Robotics iCub evolution and locomotion biorobotics self- ion assemby Started: Dec. ECCERobot theory of self- intelligence learning, development organization, neural modeling PACE robotics Amarsi 2010 self-assembly grand goal grand goal humanoid robots modular g vision Robodoc assistive roboticsp l in REAL educational (Switzerland) interfacingP prosthetics ou neural and . C L tic technology “life as it could be” y n A e EU-Cog II “life as it D th The could be” s d ro an P h ShanghAI applications to design Lectures Industrial business design art, entertainment in Lab Artists design 106 Interactive installations
- 128. NCCR: 12 year perspective EPFL University of Zurich ETH Zurich 107
- 129. follow the “Robot Companion for Citizens” initiative (“sentient machines”) FET - Future and Emerging Technologies “Flagships”: € 100 Mio per year for 10108
- 130. 109
- 131. or join The ShanghAI Lectures 109
- 132. or join The ShanghAI Lectures • global lecture series on natural and artificial intelligence • video conference with 20 universities • 3D virtual collaborative environments for classwork with over 40 universities • intercultural cooperation on interdisciplinary topic The ShanghAI Lectures, Sept to Dec 2011 109 (from the University of Zurich, Manchester
- 133. Started at Shanghai Jiao Tong University, with the generous support of 110
- 134. Participating sites 2009–2010 HamburgTallinn Michigan Edinburgh Warsaw Moscow Vermont Sheffield Salford Berlin Karlsruhe Munich Burlington Essex Zurich San Francisco St. Gallen New York Madrid Sachseln Beijing Tohoku Irvine Seoul Tokyo Tehran Xian Stanford Herzliya Osaka Chiba Algiers San Diego Shanghai Nagoya Jeddah Ain Al Taipei Singapore Sao Paulo Launceston Hobart 5
- 135. 112
- 136. or read THE 113
- 137. or read THE what book?!?? 114
- 138. Popular science Rolf Pfeifer and Josh Bongard How the body shapes the way we think — a new view of intelligence (popular science) MIT Press, 2007 Illustrations by Shun Iwasawa 115
- 139. Chinese translation Translated by Weidong Chen Shanghai Jiao Tong University and Wenwei Yu Chiba University, Japan Foreword by Lin Chen 116
- 140. How How the body the body shapes shapes translated by the way the way Koh Hosoda, Osaka University and we thinkIshiguro, Tohoku University Akio : we think : a new view to appeara new view of of soon intelligence intelligence 117
- 141. Short e-book version Designing Intelligence Why Brains Aren’t Enough Rolf Pfeifer Josh Bongard Don Berry 118
- 142. The Zurich AI Lab 119
- 143. Funding • University of Zurich, Switzerland • Swiss FNS: - From locomotion to cognition - Dynamical coupling in motor-sensory function substitution - From morphology to functionality - Swiss National Competence Center, for Research (NCCR) Robotics • EU-FET: - Locomorph - Octopus - Extended Sensory-Motor Contingencies - iCub (finished) • EU-Cognitive Systems: - ECCERobot - Amarsi - EU-Cog II/III • Private funding/others: - CIAN (Club of Intelligent Angels) 120 - Maxon Motor - Festo - Hasler Foundation
- 144. The Zurich AI Lab — Spin-offs • Neuronics • Dynamic Devices • Starmind Innovation • Enexra inc. Enexra 121
- 145. “Embodied Perception” 122
- 146. Thank you for your attention! 123
- 147. th te co eo nc re co em l “o n bo ep tica “t rc tro d og ra h l im tua l te c d e t ic nd o in s he en l a o n g tr o t al s ct n- tin s ati ry st so co ua dr ge pa on - ac e r e ft m te iv nc ce ” tc se p m e a c d en ie s sy , u s ” so hab ns one st n ft le ing nt en tu le (sk s em de no rial rs ics ch ate rgy to tron in) s r- soft robots behaviors environment sophisticated rapidly changing m or ph si c o sy m o lo de ac s nt ula -ev gy h le ce ira in et tion olu and en ga p bl m te ic /r tio c vi l is tan ity et rdi me ea n on et ron su ce hi m es ho sc th l ro of tro so l ca en l t do iplin odo bo ci lo ary log ts et g y al al ic
- 148. First hand prototype soft movements, materials Design and construction: Konstantinos 125 Dermitzakis, AILab, UZH
- 149. First hand prototype soft movements, materials Design and construction: Konstantinos 125 Dermitzakis, AILab, UZH
- 150. Wheel chair: controlled by brain waves recognition of subject’s intentions based on analysis of non-invasive EEG signals 126 Design and construction: Jose del Millan, EPFL, Switzerland
- 151. Fitness center of ten, nine, eight, … the future? Robot development by Osaka 127 University and Kokoro
- 152. Entertainment and sports ALP: The Adaptive Leg Press Design and construction: Max Lungarella and Raja Dravid 128
- 153. Entertainment and sports ALP: The Adaptive Leg Press Design and construction: Max Lungarella and Raja Dravid wichtig: “soft interaction” 129
- 154. Entertainment and sports ALP: The Adaptive Leg Press Design and construction: Max Lungarella and Raja Dravid wichtig: “soft interaction” 129
- 155. Let me be clear The brain is important!— but not the whole story “... if we want to understand how the brain contributes to consciousness, we need to look at the brain’s job in relation to the larger nonbrain body and the environment in which we find ourselves.” (Alva Noë, Out of our heads, 2009) 130
- 156. Overview: Challenges ACTUATIO CONTROL/ MANUFACT APPLICATI COMMUNIT MODELING SENSING ORCHESTR ENERGY MATERIALS N U-RING ONS IES AT-ION physical deformable artificial acquisition design metabolis functional manipulati material simulation structures muscles of control tools, m mat. on for science, methodolo changeable assembly/ soft- theory, variable gy properties surgery matter growing assembly, entertainm neuroscien compliance structures compliance model-free growth storage skin-line ent physics ce hard to actuactors model system space and self- growing/ therapy, biomechani power decentraliz identificati time density ed assembly healing human cs and bio- on resolution multi-layer self-repair assistance engineerin deposition g implement large-scale embedding morpholog building understand manufactur ical composite ing process at-ion distributed technology blocks ing life computatio engineerin embedding n underactua g organic technologi ted/ mobility ALife/AI materials? es overactuat other ed systems human/ smart comp. sensor robot programm science modalities interaction, able electrical (e.g. emotion materials engineerin 131
Editor's Notes
- #2: \n
- #3: \n
- #4: \n
- #5: \n
- #6: \n
- #7: \n
- #8: \n
- #9: \n
- #10: \n
- #11: \n
- #12: \n
- #13: \n
- #14: \n
- #15: \n
- #16: \n
- #17: \n
- #18: \n
- #19: \n
- #20: \n
- #21: \n
- #22: \n
- #23: \n
- #24: \n
- #25: \n
- #26: \n
- #27: \n
- #28: Cognition as computation: many problems, e.g.\n- perception\n- manipulation\n- movement\n- locomotion (walking, running)\nhas lead to the problems mentioned (symbol grounding, frame problem, etc.). \nIn general: neglect of interaction with real world\n\n
- #29: \n
- #30: need for an embodied approach\n
- #31: need for an embodied approach\n
- #32: \n
- #33: Cognition as computation: many problems, e.g.\n- perception\n- manipulation\n- movement\n- locomotion (walking, running)\nhas lead to the problems mentioned (symbol grounding, frame problem, etc.). \nIn general: neglect of interaction with real world\n\n
- #34: \n
- #35: always need entire organism\nhas to interact with environment in order to survive\nbody: only means of interacting with environment \n
- #36: \n
- #37: \n
- #38: \n
- #39: \n
- #40: \n
- #41: \n
- #42: \n
- #43: \n
- #44: \n
- #45: Application to human walking: forward swing of leg - largely passive. During forward swing phase: low stiffness, during impact: high stiffness.\nknee joint not directly controlled: self-organizes into proper trajectory\n
- #46: \n
- #47: \n
- #48: The fact that the passive dynamic walker has no sensors for the mechanical feedback does not imply that it’s not there!\n
- #49: \n
- #50: \n
- #51: Reflex: pressure sensors on feet —> trigger forward swing of other leg;\nbecause the proper passive dynamics is already in place, very little additional mechanism is required to achieve stable walking\n
- #52: The fact that the passive dynamic walker has no sensors for the mechanical feedback does not imply that it’s not there!\n
- #53: \n
- #54: \n
- #55: \n
- #56: \n
- #57: \n
- #58: \n
- #59: \n
- #60: \n
- #61: \n
- #62: By cooperating with the local neural circuits at the legs, cockroaches manage to climb over obstacles and to appropriately change the trajectories of the individual joints: Rather than re-calculating the trajectories of all the joints (which would be impossible, anyhow), the mechanical configuration of the shoulder joint is modified. The action of the local circuits at the legs remains the same, but because now the mechanical configuration has been changed, their effect on the behavior is different (this is speculative and it is not known whether this is actually the case).\n
- #63: \n
- #64: \n
- #65: \n
- #66: Adding sensors to make robots more intelligent is the obvious thing to do. What’s important here is that each action has a consequence in terms of sensory stimulation (see also John Dewey’s quote, below). And this is one of the fundamental differences to a computer which, in essence, “waits” for input, i.e. for someone to push a key or click a mouse button. Also: extremely impoverished sensory system.\n
- #67: grasping a bottle\n
- #68: grasping a bottle\n
- #69: illustration: drinking from a glass\nno knowledge of shape required\nforce control; passive adaptation to shape of glass\ndeformable tissue on finger tips\n
- #70: Interestingly, Dewey’s quote reads like an argument against the computer metaphor.\n
- #71: \n
- #72: \n
- #73: \n
- #74: \n
- #75: \n
- #76: \n
- #77: \n
- #78: \n
- #79: \n
- #80: \n
- #81: \n
- #82: \n
- #83: \n
- #84: \n
- #85: \n
- #86: \n
- #87: \n
- #88: \n
- #89: \n
- #90: movements of a first prototype arm\n
- #91: movements of a first prototype arm\n
- #92: movements of a first prototype arm\n
- #93: \n
- #94: Ground coffee in the pack in the supermarket is hard like a rock. Once you open it, it gets really soft. This effect can be exploited for grasping.\n
- #95: The global control parameter is the hardness of the sack - this can be manipulated through a vacuum pump: soft when balloon adapts to shape of object, hard when it has to pick it up (apply vacuum pump to increase stiffness of material). \n
- #96: \n
- #97: \n
- #98: \n
- #99: what is it about these manipulation skills that make it cheaper for Foxconn to outsource their production of the iPhone and the IPad to China rather than automating the process and produce in their home country?\n
- #100: default strategy?\n
- #101: If, with the advent of “soft robotics” the degree of automation in manufacturing can be substantially increased, this may well - according to a number of people - lead to a new industrial revolution. \n
- #102: \n
- #103: \n
- #104: \n
- #105: \n
- #106: \n
- #107: \n
- #108: \n
- #109: \n
- #110: \n
- #111: \n
- #112: \n
- #113: \n
- #114: \n
- #115: \n
- #116: \n
- #117: \n
- #118: \n
- #119: \n
- #120: \n
- #121: \n
- #122: \n
- #123: \n
- #124: \n
- #125: \n
- #126: \n
- #127: \n
- #128: \n
- #129: \n
- #130: \n
- #131: \n
- #132: \n
- #133: \n
- #134: \n
- #135: \n
- #136: \n
- #137: \n
- #138: \n