DPRG
DPRG List  



[DPRG] Hello again DPRG. (Re-Introduction) and long message.

Subject: [DPRG] Hello again DPRG. (Re-Introduction) and long message.
From: blackstag blackstag at gmail.com
Date: Tue Apr 17 14:06:03 CDT 2012

A place for both most certainly exists. I do not think ROS answers
everything nor could it ever.

I was turned onto ROS a couple years back now. At first i looked a it as
hmm I can start at the top just get stuff done. As I have gone deeper into
it i can tell you it is subsumption at its core. Simple layers fine tuned
built on top of each other with more advanced layers building on the base
of simpler layers.  Yes i know I just over simplified subsumption. Motor
control is still broken down into its normal subsumption layers and so on.
 Recent conversations with Dave and Paul made me realize this even more. I
was over simplifying what i was looking at cause those layers where hidden
from me. I did not have to worry about them, they just worked. Which is how
they should be in ROS's case. It really provides a common communication
platform for those layers to communicate with each other and arbitrate. So
the people who used it built layers upon layers. Other systems do this also
it is not alone by any means. I am not sure any one at willow garage would
agree that this is what it is maybe its is just how i view it in my limited
world. I do know once i start looking at ROS this way it became much
simpler to understand. I started to break down said tasks in my mind. The
real power of ROS is the amount of people using it and providing their hard
work back to it. It in itself is built by people with the strong desire to
solve a certain problem. Without those experts or should i say determined
people the solution as a whole would not be as complete as it is. Maybe my
understanding of Subsumption is flawed if so feel free to correct me. I
mean you Dave :) Since i know i left out arbitration and so on. I have been
doing this to long to believe i have all the answers nor do i have the
desire to have all of them. So by all means educate me ;)


/End Ramble.

Jason





On Tue, Apr 17, 2012 at 12:49 PM, Brad garton <bgart at iadfw.net> wrote:

> Thanks for the your thoughts. I guess what got me thinking about this is
> I’ve been reading a lot of posts on multiple Hobby robotics boards and
> sensed some frustration from some long term hobbyist developers, they were
> older posts by the way, that was a bit shocking to me as I was coming back
> in with a different outlook, and it caused me to pause and think about what
> my goals were.  Part of it was the feeling that I had left a project that
> was important to me unfinished. I am now and have always been interested in
> robotics, computers and programming.  Some of my lab projects in College
> were robotics applications and I caught the bug at that point.   I
> definitely still have a very positive view of the future of robotics!  This
> is just an exercise for me to set my priorities and think about how I could
> get development done more effectively. ****
>
> ** **
>
> When I originally came to DPRG, I had just left the design and development
> world and was looking for projects to do to “stay sharp”, while I worked in
> a  field job  rather than a design group. For me, robotics was the perfect
> project! I could actually SEE something happening with my code, and it was
> interesting for that if no other reason.  I soon encountered the difference
> between the concept of implementing robotics and reality that sensors and
> actuators, generally sucked in 1998.   Pretty soon my situation at work
> changed again and I was off to another place (Austin again), and MUCH
> busier than before.   The hobby languished, which is frustrating.  As a
> result I have not stayed sharp. I have let skills Atrophy that  I would
> have liked to stay current in. So I am coming back to development in many
> ways as a newbie having missed a lot of new developments the last 10 years.
> I HAVE seen what is happing in the DARPA contests, and the JPL Mars bots
> and a lot of the new developments in Academia, and it has been very
> exciting. ****
>
> ** **
>
> In 1998 (in my opinion) all that was really doable and producing good
> results in Hobby robotics was Bottom up AI (subsumption).  That was the way
> I was going then, and still am.  But there appear to be long term limits
> with that approach. As cool as emergent behavior in Subsumption is, and as
> much as it still makes sense in basic obstacle avoidance, where is the more
> intelligent behavior being developed?  The new developments out of Google
> and Stanford, would seem to suggest we need something more.  I don’t know,
> and I don’t think anyone really knows, but I do get the sense that
> something new is coming.. and soon.  Maybe ROS is part of the answer.
> Maybe integrating top-down and bottom-up is the answer. Maybe Subsumption
> still is the answer but with a lot more levels of behavior.  ROS doesn’t
> exclude that as far as I can tell.   It would be fun to find out! ****
>
> ** **
>
> It could be also as you suggest that some new idea of what life and
> intelligence are needs to come along.  We also seem to be getting closer to
> understanding biology in ways that MIGHT produce a eureka moment for  AI.
> ****
>
> ** **
>
> My problem is and I think a lot of hobbyists have the same issue that this
> is a very wide field, and it’s hard to bridge that gap between getting a
> basic platform to move, and then doing something really interesting. Maybe
> the build vs buy question for robotics platforms  is part what I need to
> come to terms with. Maybe the field is big enough to do both!  Especially
> if it’s easy to expose the things we have developed, or could easily
> develop to a standard set of interfaces like ROS. ROS and the other
> platforms like it, were something I was going to eventually go look at as
> part of what I wanted to do this time around. THANKS to DPRG for showing me
> I need to look at it sooner rather than later. ****
>
> ** **
>
> I now realize I have spent a lot longer writing, and have once again
> written an overly long post  where a shorter one probably was what is
> desired. I apologize again.   The Robotics community does get me thinking
> sometimes, and that is good!  I guess this is what happens when you quit
> lurking and start sharing. ****
>
> ** **
>
> B****
>
> *From:* dprglist-bounces at dprg.org [mailto:dprglist-bounces at dprg.org] *On
> Behalf Of *Quinones, Jose
> *Sent:* Tuesday, April 17, 2012 10:08 AM
> *To:* dprglist at dprg.org
> *Subject:* Re: [DPRG] Hello again DPRG. (Re-Introduction) and long
> message.****
>
> ** **
>
> Excellent topic!****
>
> ** **
>
> You guys have gotten me thinking (again) as to why sometimes I look at
> robotics and it almost seems depressing. Don’t take me wrong, the truth is
> that robotics have come quite a long way in the past 10 to 20 years, but I
> think the expectations were set a little bit too high (at least for me…).
> Think about the computer. 40 years ago (1970’s), and how all we could do
> was blink LED’s on the first personal computers whereas today we can strap
> to our belt a computer with cell phone, GPS, access to the internet, camera
> (for both pictures and video), game console, PDA, you name it!****
>
> ** **
>
> When I was growing up, I remember drooling on a continuous basis as I
> browsed the HEATH catalog and their HERO line of robots. It was my
> expectation that by 2000 we would have butler like robots. At the rates
> computers grew, shouldn’t that be the case? Well it is 2012 and still what
> you see mostly out there when it comes to robots are training platforms for
> kids and adults to learn about programming and robotics. Where are the life
> size bipeds ala Terminator? Sure, there are a few, mostly under research
> and development, but even then they are far from being what we thought your
> conventional Richie Boys would be able to buy in 2000. ****
>
> ** **
>
> I have also seen some quite complex platforms out there, but to deem them
> cost prohibitive is to think too positively (about 35K last time I checked,
> for just a platform with some sensors). And even then, there is really not
> that much they can do. For example, in these complex platforms I could make
> them carry drinks in a party and the robot would not collide with guests.
> But there is no way this robot could be made to actually make the drinks.
> In essence I would need to strap a second robot into the mobile robot
> platform if I wanted this kind of versatility. Could it then clean the
> dirty cups, take the garbage out, open and close the door as new guests
> show up, etc? Super simple tasks but it seems you need a single robot to
> perform each one of them.****
>
> ** **
>
> The truth is that robotics are neither simple, nor complex. What they are
> is life! That is what we are trying to accomplish here. We are the first
> life form trying to create life (albeit non biological, in this case) and
> even for that there must be a learning curve. We are a brain trying to
> create a brain, or at least a fair portion of it.****
>
> ** **
>
> I think that at the end, the successful robot will not be made of a
> collection of behavioral algorithms, but a core of algorithms that allow
> him/her to gather information and process in a way in which human-like
> learning can be achieved. At the moment, there might be computers getting
> close to doing this, but they take off entire rooms. Not to worry! The same
> happened about 100 years ago and we can put such computer about a dozen
> times in the palm of our hands, not to mention power it with batteries. The
> singularity is near!****
>
> ** **
>
> Best regards,****
>
>  ****
>
> JIQ****
>
> *Jose I Quinones*
>
> ** **
>
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org
> http://list.dprg.org/mailman/listinfo/dprglist
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.dprg.org/pipermail/dprglist/attachments/20120417/4881eb61/attachment.html 

More information about the DPRG mailing list