|
[DPRG] Re: Allen Robots (Will's emotions)
Subject: [DPRG] Re: Allen Robots (Will's emotions)
From: David P. Anderson
dpa at io.isem.smu.edu
Date: Thu Mar 8 19:29:39 CST 2007
Howdy,
Randy writes:
> > The term as I understand it is used, for example, when a
> > light-seeking behavior pulls the robot towards a goal, while
> > an IR avoidance behavior pushes the robot away from
> > obstacles, resulting in a wall following behavior that was
> > not programmed into the robot. It "emerged" from more
> > primitive behaviors acting in combination with a particular
> > environment.
>
> Just a thought, without much backing, but I would imagine, that
> if you take any of them, you have to take them all. Meaning when
> you anthropomorphize about which behaviors are emergent and
Curious choice of words.
I think what we are doing is "analyzing" which robot behaviors are
"programmed" and which are "emergent." To "anthropomorphize" is
to apply human motivations to non-humans, which I don't think
anyone is doing here.
> You have a very good example of a desireable emergent behavior,
> but I suggest the principle should apply, wether we like the
> emergent behavior or not.
Agreed.
But no reference has been made as to whether or not we "like the
behavior" or to whether an emergent behavior has "survival
value," is "desirable," "accepted," or is just annoying and we
wish it would go away.
You seem to assume that wall-following in the above example is a
desireable behavior with survival value. I make no such assumption.
Depends entirely on what the robot was trying to accomplish at the time.
Wall-following might be exactly the wrong behavior, given the situation.
But it is, in this case in any event, an "emergent" behavior.
best,
dpa
More information about the DPRG mailing list
|
