You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
NEEMs coming from both human and robot data collection often refer to body parts such as head, torso, arm, hand/gripper, finger etc. Therefore, we need to have a set of concepts for such body parts, together with axioms about how they relate to each other.
This should be mostly easy to implement, but there will be some interesting puzzles of classification. For example, we currently have Hands and Grippers as PrehensileEffectors, which is correct: both hands and grippers are typically used to grab stuff. However, what about fingers? In principle, fingers, as well as other body parts, can be used to hold objects (a complicated example of this is holding something between chin and chest).
The underlying problem highlighted by the examples above is created by a mix of desiderata:
a) a reasoning system using our ontology should be able to think about an agent's body in functional terms (what can its various parts do); this is related to a capability model, but may be a simpler stepping stone towards such a thing.
b) body parts often have a "typical function"; in the case of robots, this is explicitly so by design.
c) to some extent, body parts can be repurposed for different functions.
The text was updated successfully, but these errors were encountered:
NEEMs coming from both human and robot data collection often refer to body parts such as head, torso, arm, hand/gripper, finger etc. Therefore, we need to have a set of concepts for such body parts, together with axioms about how they relate to each other.
This should be mostly easy to implement, but there will be some interesting puzzles of classification. For example, we currently have Hands and Grippers as PrehensileEffectors, which is correct: both hands and grippers are typically used to grab stuff. However, what about fingers? In principle, fingers, as well as other body parts, can be used to hold objects (a complicated example of this is holding something between chin and chest).
The underlying problem highlighted by the examples above is created by a mix of desiderata:
a) a reasoning system using our ontology should be able to think about an agent's body in functional terms (what can its various parts do); this is related to a capability model, but may be a simpler stepping stone towards such a thing.
b) body parts often have a "typical function"; in the case of robots, this is explicitly so by design.
c) to some extent, body parts can be repurposed for different functions.
The text was updated successfully, but these errors were encountered: