Pre-built robot platforms (was Re: Let's Make Robots! site)

65 views
Skip to first unread message

Paul D. Fernhout

unread,
Jul 24, 2010, 11:11:46 AM7/24/10
to Open Manufacturing
Paul D. Fernhout wrote:
> As I think on it from a robotic dog perspective, K9 on Dr. Who is also
> related, although I don't recall ever seeing it hold a tool:
> http://en.wikipedia.org/wiki/K-9_%28Doctor_Who%29
> And Aibo:
> http://en.wikipedia.org/wiki/AIBO
> Humanities partnership with dogs has paid off well for perhaps 100,000
> years. I can hope that might be the case with robots. :-)
> http://archaeology.about.com/od/domestications/qt/dogs.htm
> "Dog history is really the history of the partnership between dogs
> (Canis lupus familiaris) and humans. That partnership is based on human
> needs for help with herding and hunting, an early alarm system, and a
> source of food in addition to the companionship many of us today know
> and love. Dogs get companionship, protection and shelter, and a reliable
> food source out of the deal. But when this partnership first occurred is
> at the moment under some controversy."

Just noticing this as I search or robot dog:
"dASA ROBOT Genibo Robot Dog"
http://www.robotshop.com/dasa-robot-genibo-robot-dog-7.html
"The dASA ROBOT Genibo Robot Dog is a Bulterrier type dog. It has been
designed in Korea where it has been extremely popular since 2005. This
Genibo Dog has a mischievous and charming personality. It does not like to
have its sides touched. It also loses energy and falls asleep when its owner
does not pay attention to it for at least five minutes. It may cause trouble
here and there, if vigilance is not kept. Your family will love your Genibo
and feel it is a real pet dog. Like a real dog, Genibo can recognize your
family and show you emotion! "

Doubt it is very open though.

Or this at the same site, which with some modifications could let me focus
on just the open software issues:
"SuperDroid HD2 SWAT / EOD Tactical Treaded Robot w / 5DOF Arm"
http://www.robotshop.com/superdroid-hd2-swat-eod-tracked-robot.html
"The SuperDroid HD2 SWAT/EOD Tactical Treaded Robot is an all-terrain
remote controlled robot designed for defense, EOD (Explosive Ordinance
Disposal), law enforcement (tactical, SWAT), security, and surveillance
applications. Built like a tank, its sealed (water tight) chassis is made
from welded aircraft grade aluminum which is ribbed and gusseted for extra
rigidity and strength. The track system is also very robust, using CNC cut
rubber treads, composite wheels with 1" thick timed teeth and Polyurethane
idler wheels which provides tolerance to changes in temperature, speed,
debris, etc. The SuperDroid HD2 SWAT/EOD Tactical Treaded Robot has the
ability to climb stairs and go over most terrain."

Although it is USD $33599, which exceeds my budget for this project by about
two or three orders of magnitude. :-) Plus I can also doubt it is open to
working on the embedded software to get it to sort LEGO automatically?

This is much more approachable, but still beyond my almost non-existent
budget at USD $3679.00:
"CoroWare CoroBot CB-LA Robot Development Platform (With Robot Arm /
Ubuntu Linux)"
http://www.robotshop.com/coroware-cb-la-corobot-2.html
"Born with a 1.5GHz PC processor, 80 GB hard disk, a 4 DOF (Degree of
Freedom) Arm, sensors and much more, CoroBot minimizes the complexity of
robotic development. The CoroWare CoroBot CB-LA Robot Development Platform
is a capable, expandable and affordable robotic platform that comes fully
assembled with an application to teleoperate right out of the box. The
teleoperation software allows the user to remotely control the robot and
read sensors. Complete source code is included."

While the hardware is not open, I'm not sure how big an issue that is, given
I'd guess it is probably legally copyable for the most part (if not
patented). So, if I wanted to focus on open software, I think that would be
acceptable to me for now. What you are mostly paying for with such a device
is convenience, as it is supposedly "Built from mostly off-the-shelf
components" according to:
http://en.wikipedia.org/wiki/CoroBot
More on it:
http://robotics.coroware.com/products
http://robotics.coroware.com/corobot

I can't say I'm 100% certain on all the open legal aspects of that, but in
general as I understand it, reverse engineering is legal unless protected by
a patent, although one can maybe argue about copyrights on certain parts
layouts or hole patterns or whatever? Anyway, great that Bryan is exploring
all that in his other recent notes.

But it looks like a neat platform to build on as far as starting from a
focus on mainly software issues, not hardware issues. I saw the same thing
on all my robot projects as well as ones by people at CMU -- I wanted to do
mostly software stuff, but just getting to the point where you had a robust
robot doing something reliably in a mechanical way could be pretty
exhausting of your time and emotion.

Again, it's perhaps not entirely open though? But it may certainly be open
enough to do some great stuff with. I'd have to add some cameras and other
sensors. Essentially, that is probably the somewhat open platform that I
could buy that would be closest to a Red Dwarf Scutter. I'm not sure the arm
turns at the base though, which is one thing that might be an issue. On the
other hand, they have a community of users, so that's a big plus:
http://robotics.coroware.com/community

There might be other platforms out there. I look over the years. LynxMotion
has also always been intriguing at the lowest prices, maybe US$1500 for a
base plus arm plus battery plus some other stuff (but a bit more work to put
together)?
http://www.lynxmotion.com/c-119-auton-combo-kit.aspx
http://www.lynxmotion.com/c-130-al5d.aspx
http://www.lynxmotion.com/images/jpg/a4wd102.jpg

But Coroware certainly is in the ballpark on price and capabilities for
someone who wanted to focus on software issues and have something working
out of the box. Their picture on that site is even of it picking up toys
(blocks). A related video of it picking up a ball:
http://www.youtube.com/watch?v=QpjRFHSr1U8

My guess is, ignoring time, to build something like a Scutter myself with an
embedded PC would cost in that range (although it might be a little bigger
or have a better computer).

Still, maybe a smaller Lynxmotion arm assembled myself on some sort of
mobile base (a repurposed Roomba) would be the absolute cheapest (although I
can wonder how strong the Roomba motors are if the thing get a bit heavy
with a battery or onboard laptop):
http://www.robotshop.com/lynxmotion-al5a-robot-arm-combo-kit-4dof-1.html
I wonder if anyone has put one of these really cheap arms on a Roomba with
some motor controller circuits?
http://www.robotshop.com/owi-535-robotic-arm-edge-3.html

On a practical basis, for a software developer and electronics person like
myself, the Coroware robot might be my best bet to build from to pick up
toys and sort LEGO, as a cheap version of a PR2. :-)

Considering how we could buy a "RoboSapien" for around US$100 a couple years
ago, and it must have a dozen motors in it, plus sensors, and how OK-ish big
RC cars are US$50, I would think within five years or less that we could see
something as capable as the Coroware robot with a couple of webcams doing
image processing for US$400 (one tenth the current price) if in quantity and
if it was plastic and maybe not quite as robust, especially as the embedded
computing side of it and wireless side of it continues to fall in price.
Spykee (from Erector, with webcam and wireless) is already less than that at
about US$200 (less on sale) although it can't do much manipulation:
http://www.amazon.com/Erector-Spykee-The-Spy-Robot/dp/B000N6470A

The robot revolution in that sense is still just beginning. Bob and Madge
(two Red Dwarf Scutters, husband and wife :-) are quite possibly in our very
near future. It might be best if their software development was open and
collaborative.

I can guess a group like WowWee might make something like that fairly soon?
http://www.wowwee.com/en/products/toys/robots/robotics
But maybe it won't be open? Still, WowWee has made their stuff open
somewhat, so we can hope:
"Programming RoboSapien & WowWee's Robots"
http://www.techtakeaway.com/articles_programrs.php
"One of the more appealing things about RoboSapien for adults, although it
is not that widely known, is that WowWee's robots can be programmed and
controlled by a computer with an Infra Red (IR) transmitter attached."

But that's not the same thing as running code on the hardware itself or
getting images back from onboard hardware that you can process.

--Paul Fernhout
http://www.pdfernhout.net/
====
The biggest challenge of the 21st century is the irony of technologies of
abundance in the hands of those thinking in terms of scarcity.

open...@gmail.com

unread,
Jul 24, 2010, 6:50:59 PM7/24/10
to openmanu...@googlegroups.com
How about building a swarm of small bots, each designed and programmed
to collect one kind of Lego?

For example one set of bots could group pieces according to weight and
others could sort them by colour and others by shape.

... you might find that simply sorting pieces by weight is good enough
to make most of your Lego collection usable again. I can't find much
info on the different weights of Lego pieces but maybe you could build
something like this:
http://www.cielec.com/high_precision_small_parts_weight_sorter.html
from cheap a digital scale with a usb interface and a couple of servos
to push pieces off the scale down shoots into different collecting
bins.

PY

> --
> You received this message because you are subscribed to the Google Groups
> "Open Manufacturing" group.
> To post to this group, send email to openmanu...@googlegroups.com.
> To unsubscribe from this group, send email to
> openmanufactur...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/openmanufacturing?hl=en.
>
>

--
Open Moto X: Open, Fast, Global.

http://www.openmotox.org

Skype: OpenMotoX
Tel: +34 96 526 9062
ES mob: +34 651 597 610
UK mob: +44 1797 224 045

Paul D. Fernhout

unread,
Jul 25, 2010, 6:50:45 AM7/25/10
to openmanu...@googlegroups.com
pa...@openmotox.com wrote:
> How about building a swarm of small bots, each designed and programmed
> to collect one kind of Lego?
>
> For example one set of bots could group pieces according to weight and
> others could sort them by colour and others by shape.

Yes, that could be possible. Although it is more batteries etc. to deal
with. Also, a Skutter sized device could put stuff on shelves or lift things
into big bins, which smaller robots could not easily do (without ramps or
something like that). But sure, simple dedicated small robots for sorting
(maybe even ones with legs or some sort of random motion, like enzymes?) is
an interesting approach. Beyond customized sensing, like with enzymes you
might have customized grippers that can grab only one type of LEGO (like a
2X6 LEGO block).

> ... you might find that simply sorting pieces by weight is good enough
> to make most of your Lego collection usable again. I can't find much
> info on the different weights of Lego pieces but maybe you could build
> something like this:
> http://www.cielec.com/high_precision_small_parts_weight_sorter.html
> from cheap a digital scale with a usb interface and a couple of servos
> to push pieces off the scale down shoots into different collecting
> bins.

Yes, that's a really good idea to use a scale. I've been thinking about that
since reading someone talking about using vibratory bowls (previously
linked) that work off of things having different weights. I agree that since
LEGO is so precisely made, weight sorting might be much easier than vision,
and probably just as effective right now. And in practice, any LEGO close in
weight probably are roughly of the same category in terms of how the get
used. Still, with vision you could look at a whole field of parts, but with
weight, you need to weight them one at a time. Still, I've thought that if
you had a big enough scale, and it was precise enough, you could throw down
a whole bunch of LEGO, have the device pick up or slide off just one part,
and then you could see the difference in weight and decide what part it is
from that difference (because that is the weight of the part you removed).

I don't know how precise the scale would need to be. I see some scales on
Amazon for around US$30 to US$50 for postal weighing (at least one with
USB), but they only seem to have about 1 gram precision at best. I'm
guessing that more precise scales are in the US$500+ range (which is a lot
more than a cheap webcam or two). Although I also now see two very small
precision electronic scales for about US$25, but maybe without USB, so there
might be some cheaper ones that are small.
http://www.amazon.com/Precision-Digital-Scale-Counting-Scales/dp/B000EG8TRG
http://www.amazon.com/Jennings-HP-100X-Precision-Digital-Calibration/dp/B001R6GA3M
And even without an interface, a webcam could then just read the numbers
that show up on the scale? :-)

I can see how some servos directly on the scale could help with that,
essentially making the scale the sorting device, similar to what you linked to.

As another approach, with the recycled scanner I just got, I'm realizing I
could also scatter LEGO on a scanner, and then just scan the LEGO from
underneath, and maybe the scanned image could be used to do the sorting? But
I agree that the weighing approach would be easier.

Ideally, one might want both weighing and visual confirmation, since the
more datapoints the better (assuming you have a way to reconcile
disagreements through sensory fusion).

Makes me wonder if one might want to integrate a scale into a gripper or
wrist on a robot somehow? Every once in a while the robot could pick up a
standard mass to recalibrate it (maybe at some specific arm angles?). Or
maybe one could use a force/torque sensor in the wrist itself for that
purpose? It's probably a patentable idea to have the robot identifying
things by how much the weigh as they actually pick them up. You would not
think something as mundane seeming as LEGO sorting would be a good research
project, but there is a first potentially patentable result. :-) Although
not now, as this is now prior art and a public disclosure. :-) So much for a
few billion dollars in patent royalty fees. :-)

Examples of such Force/Torque sensors:
http://www.ati-ia.com/products/ft/sensors.aspx
"The ATI Multi-Axis Force/Torque Sensor system measures all six components
of force and torque. It consists of a transducer, shielded high-flex cable,
and intelligent data acquisition system, Ethernet/DeviceNet interface or F/T
controller. Force/Torque sensors are used throughout industry for product
testing, robotic assembly, grinding and polishing. In research, our sensors
are used in robotic surgery, haptics, rehabilitation, neurology and many
others applications."

"Force/Torque Sensor used in a 5-degree-of-freedom system"
http://www.youtube.com/watch?v=U97qYyuUlAw

Less than I thought came up for a Google search:
http://www.google.com/#q=weighting+parts+to+identify+them
Although that scale is the first match for this:
http://www.google.com/#hl=en&q=sorting+parts+by+weight

Thanks for all the great suggestions.

Josef Davies-Coates

unread,
Jul 25, 2010, 8:08:20 AM7/25/10
to openmanu...@googlegroups.com
Only just skimmed this thread. Thought this french robot whose software was recently open sourced may be of interest:
http://www.aldebaran-robotics.com/fr

Josef.


--
You received this message because you are subscribed to the Google Groups "Open Manufacturing" group.
To post to this group, send email to openmanu...@googlegroups.com.
To unsubscribe from this group, send email to openmanufactur...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/openmanufacturing?hl=en.




--
Josef Davies-Coates
07974 88 88 95
http://uniteddiversity.com
Together We Have Everything

Josef Davies-Coates

unread,
Jul 25, 2010, 8:11:34 AM7/25/10
to openmanu...@googlegroups.com
On 25 July 2010 13:08, Josef Davies-Coates <jo...@uniteddiversity.com> wrote:
Only just skimmed this thread. Thought this french robot whose software was recently open sourced may be of interest:
http://www.aldebaran-robotics.com/fr


Apparently its the most used robot in academia in the world:
http://www.aldebaran-robotics.com/en/naoeducation

Josef.


Paul D. Fernhout

unread,
Jul 25, 2010, 9:11:50 AM7/25/10
to openmanu...@googlegroups.com
Josef Davies-Coates wrote:
> On 25 July 2010 13:08, Josef Davies-Coates <jo...@uniteddiversity.com>wrote:
>
>> Only just skimmed this thread. Thought this french robot whose software was
>> recently open sourced may be of interest:
>> http://www.aldebaran-robotics.com/fr
>
> Apparently its the most used robot in academia in the world:
> http://www.aldebaran-robotics.com/en/naoeducation

Thanks for the links, and I did not know that. That's great news.

Although the web page does qualifies it as most used "humanoid" robot, which
I could more easily believe. (Maybe from all the RoboCup Soccer matches?)
There are a lot of other smaller robots out there through the past few
decades, mostly just arms or mobile platforms. I'd expect the Roomba might
be the most used non-humanoid robot in academia at this point as it is so
accessible and cheap? But I don't know.

As to OS, I'm kind of tempted by the Willow Garage ROS developed for the
PR2, as I mentioned just now here: :-)
http://www.pdfernhout.net/open_skutter_project/

But maybe I'm just hopeful they would then dump money on me to do that? ;-)
http://www.willowgarage.com/pages/software/ros-platform

Although when I downloaded it a few months ago, I found myself a little
confused by it. Sometimes these things take a while to grok, both how they
work and the design decisions and tradeoffs and aspirations behind them.

Looks like ROS can run on the Aldebaran Nao, too:
http://www.ros.org/news/2010/03/robots-using-ros-aldebaran-nao.html
"The Aldebaran Nao is a commercially available, 60cm tall, humanoid robot
targeted at research lab and classrooms. The Nao is small, but it packs a
lot into its tiny frame: four microphones, two VGA cameras, touch sensors on
the head, infrared sensors, and more. The use of Nao with ROS has
demonstrated how quickly open-source code can enable a community to come
together around a common hardware platform. The first Nao driver for ROS was
released by Brown University's RLAB in November of 2009. This initial
release included head control, text-to-speech, basic navigation, and access
to the forehead camera. Just a couple of days later, the University of
Freiburg's Humanoid Robot Lab used Brown's Nao driver to develop new
capabilities, including torso odometry and joystick-based tele-operation.
Development didn't stop there: in December, the Humanoid Robot Lab put
together a complete ROS stack for the Nao that added IMU state, a URDF robot
model, visualization of the robot state in rviz, and more. ... Publishing
our ROS code as well as research papers is now an integral part of
disseminating our work. ROS provides the best means forward for enabling
robotics researchers to share their results and more rapidly advance the
state-of-the-art. (Chad Jenkins, Professor, Brown University)"

I can wonder if that is what is meant by the Nao has an free and open source
operating system, or if that effort lead Nao to open source their
proprietary one?

Reply all
Reply to author
Forward
0 new messages