|
|
MaximumTech
on The Robot Report
By Markkus Rovito,
MaximumTech
Tablets,
Apps, and the Kinect Bring Robotic Costs Down to Earth: Interview with The
Robot Report

Posted 08/08/2011 at 3:09pm |
by Markkus Rovito
Here at Maximum Tech, we
have a
healthy appreciation for are stupidly
gaga over robots. However, our propellerheadedness can't hold a candle to
the obsession of Frank Tobe, owner and publisher of The Robot Report and the
Everything Robotic blog. In 2008, Tobe sold his business and gave up a
successful career as a political consultant to dive headlong into robotics
research, publishing, and investing. In addition to investigating and
reporting on the developments of the global robotics industry, Tobe
maintains a robotics database of more than 1,300 robot-related companies and
has developed Robo-Stox, a proprietary method for comparing the stock
performance of the robotics industry against the Nasdaq Composite Index.
Tobe's recent article on technological breakthroughs that are bringing down
the cost of both consumer and commercial robots caught our eye. In it, he
sites among other things the Xbox Kinect as a low-cost way of adding 3D
depth sensing vision to robots (for example in the Willow Garage TurtleBot
pictured above), and tablets and apps as means for controlling robots and
making their development more open-source.
Frank opted for a little human interaction so we could interview him about
the significance of some of these recent tech breakthroughs in the world of
robotics. The following questions all derive from the original article, "Recent
Breakthroughs Are Enabling Consumer and Low-Cost Commercial Robots."
Maximum Tech (MT): You mention series elastic actuators as making
robots more safe to interact with people. What are some example tasks that a
robot like the NASA Robonaut 2 (R2) can do with its series-elastic actuators
that it otherwise could not?
Frank Tobe (FT): Safety. Imagine a robot arm swinging around from
left to right at high speed doing a task. It's all metal, gears, and
electronics, and if it hits you, you get hurt. The series elastic actuators
have an elastic spring component between the motor and the object the robot
has to pick up. The actuators help the robot detect and control the force of
its own movements. When sensors, cameras and haptics are added, the robot
can sense and stop on a dime.
Capability. R2 is more versatile than factory robots at gripping things.
Each of its humanlike fingers can hold up to five pounds, and the arm can
hold around 20 pounds in a variety of positions. The series elastic
actuators allow R2 to feel (through the use of haptic sensors and software)
the force of objects, rather than only calculate their position and make
projections about collision points. The new generation of robots coming from
the R2 project can do smaller, more sophisticated tasks, such as handling
the screws, handles, and airbag and blind-spot warning sensors that go into
car doors on an automobile assembly line. That kind of work is
"ergonomically challenging" for humans, says Marty Linn, the GM project
leader on the R2/NASA project.
Low cost. Paradoxically, the cost to manufacture robotic devices with series
elastic actuators is much less than traditional precision factory robots.
MT: What's the general
idea behind Heartland Robotics’ workplace assistant? What will it be able to
do?
FT: In addition to the massive factories that produce cars, phones
and food products, are small and medium enterprises--what are called SMEs.
There are millions of them worldwide, and they are almost all small
manufacturers without robots. SMEs have certain unique requirements: safety,
low cost, ease of training, flexibility of use, and a broad range of
capabilities. Heartland Robotics is attempting to build a flexible robot
family based on the Obrero robot from MIT, which is being redesigned to meet
the needs of SMEs. A similar project in Europe, funded by the EU, produced a
wonderfully illustrative video about SMEs titled "Coffee Break."
Heartland Robotics is still in start-up mode. Nevertheless, one can deduce
from its website the direction the company is taking: "Today's manufacturing
robots are big and stiff, unsafe for people to be around, engineered to be
precise and repeatable, not adaptable. Normal workers can't touch them...
What if ordinary people could touch robots? What if ordinary people got to
interact with them and use them? Our robots will be intuitive to use,
intelligent and highly flexible. They'll be easy to buy, train, and deploy
and will be unbelievably inexpensive. Heartland Robotics will change the
definition of how and where robots can be used, dramatically expanding the
robot marketplace."
MT: What characterizes a telepresence robot, and what are some
functions you expect to see them taking on in the near term?
FT: For one example, see the Willow Garage Texai Remote Presence
System. They are all armless personal avatars which use a web browser to
interact with and facilitate a mobile Skype-like experience, i.e., the
operator can drive the device from place to place with a two-way set of
cameras and microphones showing who is operating the device and what the
device is seeing and enabling two-way conversations. The best example of a
very necessary use was covered earlier this year on all the media (including
Good Morning America) about a bed-ridden Texas student who was able to
attend high school by the use of a telepresence robot.
The Skype-like schema for telepresence is rapidly changing with the
introduction of iPads and other tablets. All of a sudden the platform
becomes a general-purpose personal robotic device with telepresence
capabilities. iRobot's AVA telepresence robot is a platform to help robot
designers, application developers and market innovation specialists expedite
practical, affordable mobile robotic applications. It is designed to work
with a pad-based interface. The use of these tablets and inexpensive mobile
platforms has dropped the cost significantly - from prohibitive to almost
affordable. Both RoboDynamics and iRobot are shooting for a price of less
than $1,500. With a price in that range, businesses are willing to
experiment and see whether travel costs can be cut or reduced, whether
managers can efficiently carry out their duties in multiple locations,
whether technicians can coach on-site people to make the repairs that would
otherwise require an expensive visit, and whether their technical staff can
come up with other applications for the device that can solve a company
problem.
MT: Is the idea for Android apps in robotics just to use the apps to
control robots, or will it be a scenario where the app + the tablet hardware
= the robot brain, and that brain will be able to be docked on different
robot bodies?
FT: The library, tools, and hardware that come with Android devices
are well-suited for robotics. Smart phones and tablets are sophisticated
computation devices with useful sensors and great user-interaction
capabilities. Android devices can also be extended with additional sensors
and actuators thanks to the Open Accessory and Android@Home APIs. iRobot has
partnered with Google to have Android apps run on its iRobot AVA mobile
robotics platform and is open to a similar arrangement with Apple and it’s
iOS for iPads and iPhones.
Earlier in the year Google partnered with Willow Garage to get ROS to run on
Android devices. ROS is an open source and very capable robotic operating
system. As a result of porting it down to low-cost tablet devices, the
prospect of the tablet being the brains of the device could happen. But at
present the plan is that the device will have the operating system and the
tablet will have the application(s) and communications that need carrying
out. The whole concept of a $500 mobile platform with navigation system, a
$150 vision system with voice and gesture recognition capabilities, and a
$600 tablet working in sync with the these other devices, sets the
imagination afire. And that's what is happening at research facilities in
the Bay Area, Boston, Pittsburgh, Israel, in Europe and Hong Kong and Seoul:
hundreds of different apps from silly and whimsical to ingenious and
poignant.
MT: RoboEarth sounds ingenious in theory, but science fiction tells
us that robots learning beyond their initial programming inevitably spells
doom for the human race. What are some real-world outcomes stemming from
robots sharing information through RoboEarth?
FT: Robots have been incapable of coping with unstructured
environments like the ones humans work in because their systems have relied
on knowing in advance the specifics of every possible situation they might
encounter. This is one of the main reasons why robots have been relegated to
highly controlled and predictable environments like manufacturing plants.
Consequently the EU is funding a four-year program to develop a network
where robot apps can be stored, retrieved and put to use by robot operating
systems and their robots worldwide. Their goal is to enable collaboration
and sharing where little has occurred before and to encourage the
development of a shared high-level robot operating system that is not
robot-specific.
Every university and major robot manufacturer has their own proprietary
robot operating system. Few share with one another although there is some
momentum, at least within the educational world, to settle on one or two
common operating systems - such as Willow Garage's ROS. There is even
interest in and support for the RoboEarth project. But before any of the
science fiction scenarios play out, many years of development and lots of
pride and corporate protectiveness will have to be removed from the playing
field. How many years did it take IBM to embrace Linux?
MT: The last line in your article reveals that there was a recent
American government stimulus program for robotics manufacturing. That's news
to many Americans, who are told the government has no money. Can you offer
some extra information on the nature of that American government program?
FT: The case for keeping manufacturing in one's own country, manned
by their own workers and supervised by one's own skilled technicians has
been made hundreds of times and in hundreds of ways. Simply said, if we
offshore a product to a cheaper manufacturing source, we lose jobs, and soon
we'll lose skilled technicians too as the foreign resource enhances their
capabilities to include engineering, QC, packaging and design. Recently it's
come to light that many of our defense products are manufactured in Asia--an
intolerable situation.
Thus, the Obama Administration enacted and provided $430 million for the AMP
(Advanced Manufacturing Partnership) and $70 million for the National
Robotics Initiative. Much of the AMP money is earmarked to insure that
manufacturing of defense and homeland security products is facilitated in
the U.S. The money for robotics is to stimulate targeted research toward
breakthroughs that will jumpstart the kind of human/robot interaction we've
been discussing.
MT: Your sites largely cover the business of robotics, but what is
the business model for the sites -- or are they a labor of love?
FT: The business model was (and still is) three-fold: #1: to get
money from others advertising on my sites; #2: to find, select, and invest
in a basket of robotic stocks and make gains from their price appreciation;
and #3: to be open to VC and angel-type equity investments in promising
start-up companies. I've never marketed #1 beyond Google's AdSense; I'm
doing fine on #2; and I've not found a good match yet for #3. So it's turned
out to be a labor of love, but didn't intend to be.
.
The contents of all photos, graphics,
and text are protected by copyright and may not be used or reproduced
without prior express written authorization from the publisher.
|