6 S's of Capable Home Robots

Arms aren't the issue. I know Gordon has commented that the physical aspects of robots aren't progressing like computers did, but I really don't think that is the issue. A lack of adequate physical ability is not limiting the practical home robot. The ability to open doors, grab things, and roll around on wheels is already available. Even the arm is not necessary for jobs like vacuum/mop/mow.

The reason we don't have practical home robots is the lack of software that can deal with the unpredictable human world. I have a Roomba, but it really needs to learn that it shouldn't eat the sock on the floor. Of course, for the Roomba there are sensor limitations for detecting socks. But, I think the omission of a camera to see the sock is not because the camera is too expensive. Rather, it is because the software to reliably identify the sock (extension cord, carpet tassels, etc) does not exist. Once the software is created, it can copied a zillion times at low cost.

Joe Dunfee

Home computers inexpensively replicated tasks that had been proved to be possible (for others than just NASA) in mainframe computers.

Professional/commercial robotics hasn't reached the point where they can do the things we expect of a home robot yet, so hoping for home versions is too optimistic. Commercial robots have a very limited environment, and limited function within that environment.

The analogy with home computing simply doesn't hold.

Clifford Heath.

I think the basic software already exists. Look at

I have a list of 90 things that a robot could do and most of these things require arms.

I predict that within 1 to 2 years we'll have a robot like the Readybot available to hobbyist because I'm gonna be the one that makes it happen ;-)

Danh

Safety is a huge issues but I think it can be dealt with. The robot can be surrounded by a bunch of padded bump sensors that are connected to a low level safety circuit that can't be bypassed by high level code. The robot could look like a giant set of stacked inner tubes (imagine the Michelin Man). The robot would move really slow. The robot would beep, have flashing lights, ESTOP buttons, etc etc. There a bunch of other ideas I'm sure people could contribute to help with robot safety. It can all be part of a community fostering ideas about improving the system.

It's surprising that someone doesn't make a robotic vacuum cleaner which is at least smart enough to deal reliably with common floor clutter like door thresholds and cords. The Roomba/Scooba line is just too dumb to take over routine cleaning chores.

John Nagle

I'm afraid I disagree here. I think all of the hardware for building a hobby grad domestic robot exists today. Much of the mechanical hardware can be found in good ol' McMaster Carr. The electronics can be purchased from a multitude of distributors.

I will agree that a arm built out of hobby servos that can barely hold an empty soda can is not very interesting. Mind you Lynxmotion will sell you one that can lift up to 13-oz.

Far more likely that some hobbyists will just build one to show off to their friends.

I never said anything about safe or reliable. I think the initial hobbyist grade domestic robots will be neither safe nor reliable. If you have small children, keep them away. The first ones will be about as reliable and safe as a model-T Ford.

The hobbyist could care less if their robot is kid safe. They can keep the kids away from the robot separated for safety.

Any parent that exposes the kid to an unsafe conditions, should have their child custody privileges removed.

I think a commercially successful domestic robot needs to be safe and reliable, but that comes later.

-Wayne

And the word is "inexpensively." You nailed it on the head, Clifford.

Sure, the software and hardware for a home robot probably does exist, though there seems to be some disagreement here. But if it does exist, it does not exist at a price that makes it practical. If you can't afford to build it, it might as well not exist. It's out of your reach one way or another.

The battle bot TV programs showed perfectly what you could do with enough money. On one program (forgot which one) there was some hydraulically operated referee bot that was 2-3X the size of the others, and probably cost a small fortune to make. No contestant robot could take it on. With enough money you can do just about anything.

In past messages I've brought up the notion of business and economics for a reason.

Also remember that the lack of a product of an obvious idea doesn't mean everyone else has overlooked something. If a functional (key word here) home robot doesn't exist, despite there supposidly being the technology for it, there must be a reason. What could that reason be?

-- Gordon

Yes, and it's embarrassing. Robotics competitions in Japan are way ahead of that. See "

The toy industry is doing better. Look at WowWee.

John Nagle

Agreed. However, the IEEE annual rate of $169 plus $21 for a Robotics and Automation Society membership, can be a little steep for the hobbyist.

If you have access to technical university library, you might be able to read them by visiting the campus library.

-Wayne

Well, that wasn't my point. Quite the opposite actually. A home robot has to be able to do fifty different useful things in a fairly unconstrained environment. Not even NASA has built a robot that's that capable, let alone industry. Until the industrial capability exists, it cannot be made inexpensive.

" Also remember that the lack of a product of an obvious idea doesn't mean everyone else has overlooked something. If a functional (key word here) home robot doesn't exist, despite there supposidly being the technology for it, there must be a reason. What could that reason be?"

That's a million dollar question but it happened with the Roomba. The technology in the Roomba was available back in the 80's but for whatever reason it took 20 years for a company like iRobot to come out with an affordable robotic vacuum.

I don't believe at the price needed to make a commercially product viable, and that includes a price low enough to encourage the average consumer to buy it. At (say) $200 they have to make the Roomba for around $50, at the most. That's a lot of robot for fifty bucks. The Roomba is also quite deceptive. Robots in the 80s could barely be relied upon to re-home; it's rare when the Roomba can't.

People are still forgetting a basic tenet here: with enough money you can build just about anything - a rocket to take you into outer space. But can you afford it? Just because it's technologically possible doesn't mean is feasible.

-- Gordon

I think only to you because you're not thinking as a new learner. Kids still need a starting point, and establishing a fundamental baseline is a principle in any educational endeavor.

Japan may or may not have an edge, but until recently their economy was in better shape than ours and they could do more with their money. Things change and priorities reset. The US has lagged FAR behind many countries in allocation of funds for education. What you're seeing has much more to do with the lower value American taxpayers give to educational goals.

-- Gordon

Is a $2,000 Readybot like robot considered affordable? To me something that is not affordable and a bit out there is all the full size humanoid robots coming out of Japan (ie: ASIMO). I think a PC on wheels with arms is doable and affordable with today's technology.

I thought we were in agreement here, but maybe I misunderstood your comment.

NASA **could** build a domestic robot of the type needed to open stuck pickle jars, get milk out of the refrigerator, and pick up the loose kitty food off the floor. They have chosen not to. A research lab of the caliber of an MIT or Carnegie Mellon could do the same, and probably some gifted individuals with enough time and money on their hands could as well. Obviously, all the above have chosen a different path, or else they've kept their developments secret.

-- Gordon

Not affordable for what I saw Readybot could do as a domestic robot. As a research or learning tool that's quite a different matter.

A robot that dumps food from a plate into the sink (especially when the plate needs to be set out so that the robot can get to it) is not a compelling story, IMO. (Yes, I know it's just a demonstration, but it doesn't move me in any way.)

A robotic device that swims in the water of a backyard swimming pool, senses the erratic motion of when a person falls in and is flailing about, and attempts to rescue that person by going over and becoming a floatation device, now that's compelling. Especially compelling if the robot is also cleaning the pool as it swims around.

-- Gordon

I'll add "rescue drowning person from the pool" to my list of 90 things a robot could do.

In terms of saving a life, a capable home robot could sense smoke in a home and open the back door to let the dog out into a fenced back yard. Rescuing one's dog from a fire would be priceless to most pet owners. A robot could also let the paramedics into a home if it sensed that an owner had collapsed due to a life threatening situation (ie: heart attach, diabetic shock, etc).

Maybe. On the face of it, it sounds like a good plan to open doors in a fire. Unless you've seen Backdraft. Oddly enough, it's probably safest not to have robots go around opening doors unless they have all sorts of sensors to determine if it's safe. But besides that, it would be far cheaper - and probably more reliable (speed, obstacles, whatever0 - to just make a back door that automatically opened. With backdraft sensor and automatic lock maybe $200? Dunno, but it sounds more reasonable and cost-effective than a do-all robot.

-- Gordon

No, I don't believe they could, to a quality suitable for industry. One in twenty times, the pickle jar would get broken, the milk dropped, and the kitty litter smudged into the carpet. They'd have a manual override that could activate the recovery of such situations, as they did whenever the Mars rovers got stuck, but that wouldn't be acceptable in either a home nor an industrial environment. Just think of the liability claims, for one thing.

Until such things are profitable in an industrial environment, where the payoff may be 10-100x, they won't become cheap enough in the home environment.

Clifford Heath.