Nice list, but it is much easier said than done. I do disagree with
"There is absolutely no reason why we can't have capable robot in our
homes TODAY. "
Many have tried, but it is certainly proving harder than it was first
thought. On the drive to work, I've been listening to Podcasts of old
radio broadcasts of science fiction stories from the 40's and 50's.
They really were expecting us to have farms on mars, and robots doing
all our housework by the 80's.
I suspect that, for the near future, the only things we can really
hope for, will be a practical robot vacuum and a lawn mower. This just
requires a bit more capability than current similar products; things
like more navigational certainty, and visual processing to avoid
obstacles. Autonomous robot airplanes for the military are relatively
easy, since there is nothing for them to bump into.
Although true autonomous robots are a problem, I suspect there will be
significant advancement in remote control ore teleprocessing
"robots". This is where there are "brains" in a machine that help
prevent it from driving off a step, but it will ultimately be operated
by a much more advanced processor that God designed and placed in the
skulls of human operators.
I accidentally hit reply to author instead of replying to the group so
here's take two.
On Feb 22, 1:37 am, firstname.lastname@example.org wrote:
I can see most people disagreeing with the above statement since it
hasn't been realized yet. The only way to back up the statement is for
me to build it but that's the great thing about robotics.
I see a lot of technology today that could really make these bots
real. WIFI, the internet, cheap computers, cheap actuators, wireless
video, webcams, USB, micros, software IDEs, etc, etc, etc
here are two companies that make vision products:
I've used the stuff from Evolution and it is good enough to recognize
objects in a fridge and pantry.
The PR2 from www.willowgarage.com takes the approach of teleoperation
and vision processing. I'm kinda building a poor man's version of that
bot. I do agree that ultimately a human operator needs to have the
Someone gave a lecture at CHAOS this year on this sort of subject.
Their opinion was that insurance is the test. If you cannot insure a
device that works autonomously, you will (probably) not be allowed to
use it. The insurance companies have not enough data yet to quote.
Someone has a robotic lawnmower (self-made) working in their garden,
a stone hits one of the control boxes (perhaps by bouncing off one of
the front tyres, not an unknown instance, even in human-controlled
ones), vehicle goes haywire, gets out onto the public road, and is
hit by, or hits, a car, or human.
(The above was solved in one real-life instance by the operator
running after the device and turning it on its back. If you have to
be in attendence the whole time, whats the point of having a robot?)
That's an interesting perspective. Suppose we replaced the autonomous
lawnmower scenario with an owner's dog. If the dog runs into the
street and causes an accident I wonder if it would be covered by home
owners or auto insurance? Maybe neither???
Probably depends on the state ("no fault" car insurance) but if your dog
bites someone, having homeowner's insurance sure would come in handy. If
you don't have that you might be in a lot of trouble.
The lack of liability insurance is definitely why we don't see 200 pound
robots roaming through the house doing the dishes or cleaning up after
the cats. The iRobot is specifically lightweight so that it won't hurt
toes or animals or fancy furniture.
Safety training would be a requirement for owners of big robots. There
are also ways to minimize risks by using distributive robotics. In one
scenario a big robot with arms may be in a confined area and it would
avoid humans (call it the shy robot). The shy robot could grab objects
out of a fridge and then transfer those objects to a smaller/less
powerful "butler" robot that could interface directly to people.
It still troubles me that I see robotic "experts" disregarding safety.
Here's a video of MIT folks showing their socially interactive robot
called Ripley to Allen Alda. It's disturbing how the robot is behaving
unpredictably yet the researchers allow it to get so close to Allen's
I'm having difficulty understanding why a robot is needed for this. I
know it's just an example, but I don't think a clear case for a domestic
robot can be made without first defining its role. Personal computers in
homes didn't really take off until applications such as Visicalc were
introduced; before then, they were strictly for the hobbyists.
No one is going to pay for a domestic robot that gets things out of the
fridge. That's the sort of thing we robot builders do as an experiment,
but it's not viable business.
My opinion is another fifteen years in industry, another ten after
that in commercial applications, then a very slow expensive creep into
domestic. I build near full size bipeds and there is NO way mine roam
unattended. Mechanics and brainpower are just not even remotely close,
and sensor fusion and interface has Sooooooooooo far to go.
I have had pretty good success in my small shop getting things
working, but when things do go bad-It is BAD. Even machine less than
100 lbs make a really bad mess when they go down(notice I didnt say
if). I have had more than a few instances this week. Just when you
start gaining some confidence, thats when you get smacked in the
Fetching objects from a fridge could be helpful to people who have
difficulty performing activities of daily living (ADL) such as
walking. It falls into the category of an assistive technology.
Here's a list I started 2 years ago for a class listing 90 things a
robot might do in the home. Anybody want to add to it? - don't think
too hard, just brainstorm.
While I agree with the basic premise, I don't think it's lack of
leadership because coming up with a truly useful home robot that would
be affordable would likely cost billions to develop, and the return is
far from a known. In Japan, where they are closest to having domestic
robots, to pay for their R&D they have instead turned to specific
applications like elder care (homes, hospitals, nursing centers) because
one robot could be shared or reasonably leased in order to spread out
the costs. The environment is also more controlled - a robot is bound to
be able to roll over the same flooring as an aged person in a walker.
Look at how long it's taken for someone to come out with a widely
available set top PVR that also connects to the Internet for VOD,
YouTube, and so on - the true bridge between a PC and a cable box. (What
they have now is limited in market, capabilities, or both.) If it takes
cable and satellite companies this long to figure out something that
people would actually want to buy and have an immediate use for, I can't
see that domestic robots would fare much better.
Great domain name, BTW!
I'm a bit more optimistic about home robots. I'd like to point
The robot in videos (click on [Resouces] under Main Menu) costs
about $9K. It is using some very expensive electronics that can
be substantially cost reduced. Indeed, I think the main robot
can be built for about $1.5K. (I'm not making this up, I've spoken
with the ReadyBot team.)
I fully expect the first home robots will be for hobbyists and
over time the market will expand from there. The personal computer
market took several decades to reach the current confusing state of
affairs and I expect home robots to follow a similarly tortuous
Gordon McComb wrote:
Here's some more Readybot video:
Readybot at last nights HBRC meeting:
(Readybot starts about 3-4 minutes in.)
Readybot presentation from last month's HBRC meeting:
(Starts at about 10 minutes in.)
I had seen this recently and I thought it was a terrific 'bot.
Don't get me wrong: I'm not against the idea, but I feel the major
stumbling block is presenting a good business case for a domestic robot.
Even at $1,500 it's on the high end for hobbyists; there is simply no
way the general public would pay $5K-10K that would represent the
starting price for these - considering R&D, manufacturing, marketing,
licensing (things like proven vision systems, and so on), liability
insurance, and dealer markup.
I recall the Whitebox robot was going to come out under $2K. The panels
alone cost about that. Things get expensive when A) the market is
relatively small and B) the product has limited appeal and C) the tech
pieces aren't quite there yet. For example, what mods would ReadyBot
have to have in order to open a can of dog food? I guess a can opener
attachment, then there's a question of what it needs to open a glass jar
of baby food and another a large jar of pickles, and so on.
I view these kinds of products as germination for further research, but
I question whether there's yet a market that could sustain itself.
Now, if you want to talk about burgeoning markets for robotics, there's
a whole slew of teleoperated apps that are waiting to be tapped, bigger
and better (and badder) sentry robots (probably less liability involved
than guard dogs), human-interface assistants (i.e. Keepon, the subject
of another thread here), and many others.
When you talk about prices putting things out of reach everyone should
remember what a Macs (and PC's) cost in the 80's. Seems to me they
went for well over $5000 and seemed to sell. So what does $5000 then
translate into todays costs? Computers were after all a "niche" item
back then as robots are now. Toys sell for under a $1000, something
usefull and reliable can sell for much more. Time and acceptance will
bring the cost down so don't let price be such a scare point in your
development. Build/design something usefull, keep the cost down if
possible, but make it reliable first, cheap second.
my 2cents worth
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.