La lère conférence Humanoïds2000 s'est tenue les 7/8 septembre 2000, au MIT. http://humanoids.uni-bielefeld.de/home.html

Elle était sponsorisée conjointement par l'IEEE et la Robotics and Automation Society  aux USA, ainsi que par la Robotics Society of Japan

IEEE: Institute of electrical and electronics engineers http://www.ieee.org/
IEEE Robotics and Automation Society http://www.ncsu.edu/IEEE-RAS/
Robotics Society of Japan http://wwwsoc.nacsis.ac.jp/rsj/

Les organisateurs d'Humanoïds2000 précisent les objectifs suivants:
Due to recent developments in enabling technologies (processing power, mechatronics, walking machines, articulated vision heads and more) but also due to findings and developments in other fields (e.g. studies of the human brain, linguistics, psychology), we currently observe a shift in the view of what is artificial intelligence and how it can be put to work in operational autonomous systems. This sets the stage for putting perceptive, cognitive and manipulatory abilities together to create truly autonomous humanoid robots. Unlike autonomous service robots that perform a more or less limited range of special tasks without human supervision, the autonomous humanoid combines advanced manipulation skills with human-like cognitive processes embodied in (approximate) human shape so as to be able to operate in unchanged man-made environments. It receives its tasks by carrying on a dialogue with human instructors involving speech, gestures and facial expressions. And it can use the same tools and appliances as human beings. A number of impressive mechanical platforms have been designed over the last decades, but, with very few exceptions, they have not been coupled with "brains".

This is the reason, therefore, that we feel the time is ripe for bringing together researchers from different fields that share a common interest in this intriguing new field of interdisciplinary research. Since we intend to establish a conference series, the preponderance of papers of this first conference should be on identifying current research trends, presenting and reviewing recent work and, possibly, speculating about the future. The contributions should address the general question of what the roadblocks towards full humanoid autonomy are and how they can be removed. They should also be clearly related to the general theme of the conference, i.e. it should be stated how and why the work presented may be applied to present or future autonomous humanoids.

Les 4 domaines d'études ont été les suivants:

1. Brains, Mind and Robot Epistemology
Philosophical implications of artificial embodied minds
Embodiment as the basis for intelligence: co-evolution of cognitive and manipulative skills along with internal representations
Autonomous construction of categories and concepts through interaction
Grounding of behaviours and reasoning in sensory patterns
Higher-level representations and consciousness
Mental simulation, emulation and planning
Anticipation as a hallmark of intelligence
Learning and sequencing of high-level behaviours
Coupling of sensing, behaviour and reasoning (situated perception action cycle)
Fault tolerance through multisensor perception and compensation
Issues of self-reproduction
Neurophysiological findings as a guide to artificial brain design

2. Interaction with Humans and the Environment
Control and instruction through dialogues: recognition, production and integration of natural language, gestures, facial expressions
Development of language and gestures through observation and imitation
Desires, intentions and emotions and their expression
Simulation and rendering of face expressions
Keeping the focus of attention; recognizing specific situations
Body motion as an aid for the humanoid to express its internal state (of mind)
Communication between humanoids (human-understandable vs. proprietary)
Human psychology, social acceptance of humanoids and culture/country-related issues
Intermediate applications and habitats (toys, help for the elderly, deep space missions, ...)

3. Structure and Purpose of Body and Limb Movement
Body gross and fine motion and its learning through demonstration
Learning of grasping and sensorimotor control for complex manipulation tasks; adaptation to structurally new tasks
Physical interaction with human-beings or with other humainoids for cooperative tasks (with or without explicit instruction/communication)
Navigation, planning of movement and collision avoidance
Body-centered behaviours (limb coordination), posture stability
Locomotion, gait and foot-placement
Simulation of all aspects of body and limb dynamics

4. System and Components Design
Mechatronics of hands, feet, legs, arms, heads; new types of actuators
Materials for actuators and skin
Sensor design (articulated vision, tactile, directional ears, new sensors)
Evolutionary hardware and control software
High level programming
Architectures for component coordination and distributed control (multi-agent-systems, distributed control, organizing internal multi-processor systems)
Power-supply
Building an international infrastructure for humanoids research (simulation software, shared use of components or full robots, ...)
Suggestions for future contests for full-fledged Humanoids or their components: ”HumanoCup”

Le bâtiment principal du MIT © MIT

Programme détaillé des interventions: http://humanoids.uni-bielefeld.de/program.html Nous n'avons qu'une chose à ajouter: la recherche européenne a de quoi faire pour ne pas prendre de retard. C'est un veritable déferlement américano-japonais qui se produit.

© Automates Intelligents 2001