Howdy,
Following on from the behavior based robotics discussion, it seems clear why a researcher into artificial intelligence might view robotics as a distraction, even irrelevant. If the main topic of interest is AI, then robotics is certainly a minor player.
However, for those of us who are interested in robotics as a field unto itself, AI is a natural component of robotics.
That is, that robot designers/builders need some sort of guidance on how to organize and order their software based on the robots we are building now, not hypothetical robots of the future with human-like intelligence and the ability to learn.
of AI has made available to robot builders is the subsumption architecture, or so-called "behavior based robotics," in all its many variations and dialects.
There seems to be an emerging consenus on this list that this approach has "stalled," evidently because of the paucity of new papers being published on the subject. My experience is that this is not so, but that may be because, like everyone else, I use a modified, even bastardized, concept of BBR, for which I do no see any particular plateau approaching.
We've heard a lot of statements like "the BBR approach cannot solve complex problem sets" with no actual experience or proof to back it up. Like the fellow who wanted to close the patent office in the 1800s because "everything had already been invented." At least everything he could think of. My experience is otherwise.
Perhaps this perceived plateau is actually among the theorists and their debating societies, and not among actual robot builders and their creations? That seems to ring more true.
best dpa