|
[DPRG] Re: Allen Robots (Will's emotions)
Subject: [DPRG] Re: Allen Robots (Will's emotions)
From: R. Steven Rainwater
srainwater at ncc.com
Date: Wed Mar 7 10:47:12 CST 2007
On Tue, 2007-03-06 at 02:02, David M Wilson wrote:
> In the current implementation the focus has been on
> developing proportional responses to stimuli. Not
> just for steering control, but for braking, for
> excitement, etc.
That's very much what subsumption is about too. It's a basically a
reactive, non-representational, behaviour mechanism similar to the
mechanisms found in simple organisms that lack a brain such as protozoa
or nematodes. There are other reactive, non-representational behavior
mechanisms out there too, but subsumption has so far been the most
successful at making real world robots do interesting (dBot) and useful
(Roomba) things, probably because it emulates a solution that evolution
has tested and refined over millions of years.
> one measure of success for behavior based design is the
> appearance of organic-like behaviors that are not intentionally
> coded.
I'd agree with that. This is known as emergent behavior and seeing it in
a robot probably means you're on the right track.
> What is the point in seeking organic behaviors from a device
> that utilizes odometrics for navigation? At best you'll get
> canned and well rehearsed demonstrations.
The use of odometery is known to be present in biological organisms. A
good example is the Sahara ant, which uses a combination of step
counting and current heading to orient and localize itself. The current
heading is determined through the use of polorized light sensors that
determine the sun's orientation. The sun's position is compared to an
internal ephemeris table to determine which way the ant is facing. (more
details on Ant Odometery here: http://robots.net/article/1968.html )
On the other hands, American ants mostly just follow pheromone trails
left by themselves or other ants. The Sahara ants evolved fancier
behavior because pheromone trails don't work in the desert.
Both types of ants show similar "organic" behaviors, so I think this
means the underlying navigation system is less of a determining factor
than higher-level behaviors.
-Steve
More information about the DPRG mailing list
|
