[ros-users] Current state of SMACH in ROS

Ingo Lütkebohle iluetkeb at gmail.com
Tue Feb 21 09:16:41 UTC 2012


I guess the remaining questions should, as suggested. really better be
talked about in person, with a few examples handy. However, I made a
mistake in my last e-mail, that I just wanted to correct here.

On Mon, Feb 20, 2012 at 2:34 PM, Herman Bruyninckx
<Herman.Bruyninckx at mech.kuleuven.be> wrote:
>> For me, a behavior is the combination of several tasks to achive a
>> goal.
>
> Ok, I see. Not my semantics, really, but mine is biased by a history in
> control theory and dynamical systems, and less in "AI" :-)

In fact, I was not so much AI inspired, but just plain wrong ;-) My
definition above is that for a "plan", not for a behavior. A behavior
is certainly much closer to your definition, in AI, too.

That said. it captured what I care about with tools such as SMACH:
Executing plans, that is, a (often hierarchical, sometimes cyclic)
sequence of steps towards a goal. In my case, most of the plans are
not actually generated at runtime by a planner, but pre-specified, but
nevertheless.

The term "sequencer", which was mentioned by Alex, is not the one I
usually use, but it captures it quite well. Part of the work of a
sequencer may entail coordination, but its really not the same, and
maybe a sequencer should not also be a coordinator, I can't really
tell right now.

Lets just leave it at that, I guess ;-)

cheers,

-- 
Ingo Lütkebohle
Bielefeld University
http://www.techfak.uni-bielefeld.de/~iluetkeb/

PGP Fingerprint 3187 4DEC 47E6 1B1E 6F4F  57D4 CD90 C164 34AD CE5B



More information about the ros-users mailing list