Home
Robot Killer App
Apr 29, 2013

In an earlier post I asked which topics you prefer me to blog about. I was surprised that the topic of robots wasn’t popular. I say that because interest in anything is usually based on how much we expect it to influence our own lives. And I can’t imagine anything that will change all of our lives as much as the coming Robot Age. We are the last generation that will remember life before robots. In about five years, shit is gonna get real.

So how will the first big push into residential robotics happen? If you break your daily chores into categories, which of those categories do you see as the first killer apps for robots?

We already have robots that vacuum carpets. But armless robots that scoot along the carpet don’t impress me. I’m wondering when I’ll buy the first robot that can move through my home, manipulate things with its arms, and communicate by voice.

The other day, I was clearing the dinner table and putting dishes in the dishwasher. It occurred to me that technology has already reached the point at which a robot could clean your kitchen and dining table after a meal. A robot could collect plates, scrape the debris into the trash, and load a dishwasher. If you saw the Youtube video of a robot doing ironing, or pouring a glass of water, you know why I think the technology is already here.

I could also imagine a robot walking the family dog when everyone else is working or in school. You’d need the robot to train the dog with treats, just as a human would. But that seems doable. And the robot would need to have a DVR function in the cloud to record whatever is happening and discourage pranksters and dog thieves. Perhaps you, as the dog owner, could watch the entire walk on your smartphone or computer at work. You could even talk to passersby through the robot, just so they know they are being watched and video-recorded.

I also have a vision of a Transformers-like robot designed specifically for childcare. At night it turns into a bassinet with video feeds to the parents. If the baby cries, the parents can have the robot bring the baby to their room, or try to soothe it by rocking. When it’s time for a walk, the bassinet becomes a self-moving baby carriage. For meals, the robot morphs into a high stool. If the smoke detectors in the house go off, the robot carries the baby to safety automatically. And I would think it could sniff out a diaper problem and alert parents. A robot won’t replace adults for childcare anytime soon, but I can see childcare getting a lot easier with a robot helper. In time, the robot can even teach the kid language skills.

I think it will be a while before robots can cook gourmet meals or clean your bathroom just right. But I’ll bet we’ll have commercial robots that can clear a dinner table, take the dishes to the dishwasher, do laundry, and help with childcare in five years. And I think the price tag will be around $5,000, with a monthly maintenance plan of $100.

The first company that cracks the residential robot market has a good chance of becoming the most important company on earth. The robot revolution will make the industrial revolution look like practice swings.

By the way, if there are any college robotics majors looking for a fun project, I have one for you. I need a small robot that can find and pick up tennis balls from anywhere on the court and throw them in a hopper (a basket) on its back.

When a human takes a tennis lesson, or uses a ball machine to practice, the unpleasant part of the process is picking up the two hundred balls that are left all over the court. A tennis teacher could save ten minutes of tedium from every hour-long lesson.

I would think the technology for a tennis ball robot is already here. Let me know if someone already built one.

 
Rank Up Rank Down Votes:  +25
  • Print
  • Share
  • Share:

Comments

Sort By:
Apr 30, 2013
"But armless robots that scoot along the carpet don’t impress me. I’m wondering when I’ll buy the first robot that can move through my home, manipulate things with its arms, and communicate by voice."

I'm sure that a hundred years ago somebody said to the Wright brothers, "I will be impressed only when you build a machine that can fly by flapping its wings."

Can't see why you're imposing human limitations on robots.
 
 
-1 Rank Up Rank Down
Apr 30, 2013
Speak and Spell taught me everything i know about spellng.

For my robot vision of the future, one word: Plantation.
 
 
Apr 30, 2013
I think you are dead on, Scott. Big changes in technology often APPEAR at first to be simple geometric progressions but once you get a little further in you realize they are actually exponential. Look at cell phones. Initial growth, and then KAPOW. Nearly everyone on Earth has one. The same was true with home or workplace computers.

It's hard to say if tablet computers are there yet since it is still early in their life, but the workplace aside (a big hand wave admittedly), the desktop is becoming a rounding error.

I agree that robots will be transformative in the same way the cell phone and desktop computer was.

Jobs I encountered today that *will* be placed by robots:
most of the car drivers I commute next to (I drive a Vespa)
The cashier and barista at Starbucks.
The person who cut my hair (if a robot can paint a car perfectly, it can certainly cut hair to most people's satisfaction, we are not unique snowflakes).
The UPS driver and delivery guy (half of this job is already roboticized)
The telemarketer who called me at dinner time.

 
 
Apr 29, 2013
I work with industrial robots, and that experience tells me the hardest part of a feasable home robot would be programing. Home Robots would need some sort of AI to constantly reprogram themselves "on the fly" for various obstacles to their standard routines. Daily radom things such as a sleeping dog, kid's toys ect. would have to negotiated constantly. While simple tasks such as vacuuming simply need avoidence procedures more complicated tasks (carrying a baby) would require multiple & complicated sub-routines to avoid any harm to the child. I know how much time & skill it takes to progam an industrial robot to perform just (1) complicated routine (that never varies). Programing a home robot with the level of interaction you wish for would require an extrodinary amount of time, and would be good for a single eviroment only. I believe we are a good 20 to 30 years away from robots the way you envision them. I also think cost ans litigation issues will stop them from being anything but toys for the wealthy.
 
 
Apr 29, 2013
Sexbots.
 
 
Apr 29, 2013
It's easy to imagine a reasonable $5000 hardware cost.

You'll probably need no more than 15 of ~$15 servo motors for fingers and other fine motor controls. Another 20 (max) high-torque $50 servos for the heavier-duty joints. Add in a $100 quad-core CPU board, maybe 3 controller board in the $40 range to communicate between CPU and Servos, and a pair of VGA-quality webcams for computer vision at $10 each. You probably want tilt sensors to help with balance, that's another $10. $100 - $200 for batteries. The rest of the robot is < $500 in plastic, and no more than $1000 in labor to assemble a decently design machine.

Add some fudge, that's $3000. For around $2000 profit on a $5000 android, and I'm using retail prices on this stuff -- should be cheaper for a manufacturer buying thousands of parts, and I've left lots of wiggle-room.

Really, with high-power CPUs, memory, storage, and controller interface cards at ridiculously low prices, the real challenge is software. Once it's made, it's near-free to duplicate.
 
 
-3 Rank Up Rank Down
Apr 29, 2013
Robot: Stop...peeing...on...me...stupid...dog...I'm...rusting.
 
 
Apr 29, 2013
No robot posts because your robot posts are ... optimistic?

I'm not sure that's the right word.

Horribly optimistic?

Better.

Five years is far too short. Sure, the tech is there, but home-bots need to have safeguards to avoid crippling lawsuits. Why do you think Roomba's are still little discs that shuffle around the house? It's certainly not because the engineers at Roomba can't figure out how to make them climb stairs. It's because once you make them more autonomous and more capable, you make them more dangerous - more able to hurt themselves and others. The inherent liability of selling these things to stupid rich people would be enormous. We do not yet have the tech to make sure idiots, their children, and their pets are safe with robots more mobile than a Roomba.

Once we do have proven safety tech for a more mobile and more capable robot I'd go along with a 5 year time horizon, assuming a reasonable price and cost of ownership. We're just not there yet. Maybe 10 to 15 years from now we can start the 5 year clock ticking.
 
 
+2 Rank Up Rank Down
Apr 29, 2013
I'm comfortably certain Scott uses a maid to vacuum his house.
 
 
+4 Rank Up Rank Down
Apr 29, 2013
Another area where robots will have a huge impact is driving. Think of how many people are employed as drivers (taxis, trucks, buses, chauffeurs, shuttles etc.) Now think of how many people are employed because they and the drunks they share the road with frequently screw up (cops, tow truck drivers, EMTs, lawyers, insurance adjusters, ER doctors, medical staff, etc.). Google has already shown us drivers can be replaced, it's just a matter of sorting out the liability.

See Marshall Brain's essays on the topic. His prediction: Airline pilots are the first to go.
http://marshallbrain.com/robotic-nation.htm
 
 
Apr 29, 2013
This is really an economic question of the trade-off between Labor and Capital. It is highly likely that we will continue to substitute Labor for Capital, especially in the service sector, as we have mostly done all we can in manufacturing and agriculture, though the capital will become even more efficient. It is never a 100% substitution though.

The only other question is if people have some preference for capital(robots) that is not based on efficiency(i.e. cost minimization). You might be willing to pay more for a robot servant, that is as productive as a human servant. For example, it is a matter of market "repugnance" to hire someone to do certain things, even though there may be willing parties on both ends.

As for the voice communication, I feel this is a misplaced hope. Voice communication is just naturally inferior in communicating information. Your best tool for communicating information is your fingers, be it a keyboard or some similar yet-unrealized input system. Note I also put little stock in brain-wave readers.
 
 
-3 Rank Up Rank Down
Apr 29, 2013
BTW...I would like to have a company like a "US Robotics" that I can buy robots from (ala Isaac Asimov) but it is far into the future and not the 5 years as discussed.
 
 
-2 Rank Up Rank Down
Apr 29, 2013
I'm not so sure you can achieve the price point of $5,000 per robot. Even with PC's which is at a mature phase of its life cycle, you can still spend anywhere from $700 to $2000. Add on the manufacturing costs of actuators, gyros, long-life power sources and even at a mass-production scale, the costs go through the roof. My guess is it would easily cost $25,000 to $40,000 for a mass-produced robot that is just able to move around the room on its own without knocking over stuff. If you have to program it move around your house and furniture and remotely command its actions then it really isn't an autonomous robot is it?

Having it be able to do anything more than merely walking a dog would add onto that cost. It cost millions for the recent DARPA robot that is able to just drive around a set off-road course without running into things or getting itself into trouble. So the argument about how much is this thing going to cost for it to appeal to consumers is going to depend on whether you are talking about a direct command/programmed piece of machinery or a true autonomous robot able to understand your voice commands as an interface and react to its environment based upon your directions (i.e. clean the windows on the first floor).
 
 
Apr 29, 2013
Scott ... you are probably right but I sincerely hope the 5 year prediction turns out to be 25. All we need right now is something else to make us even more sedentary than we already are. I know, you're thinking the robot will free our time so we can participate in productive exercise, like playing tennis. I think we have made life "easier" enough for now, and of course by that I mean "easier" for the wealthiest 20% of the world that can afford it. The other 80% of the "poor folk" still have to scramble to get enough to eat and they are thinner and probably healthier for it, at least until they get malaria or dysentery. Hmmm, maybe your idea for a robot that can analyze health issues, and maybe even that same robot could detect dirty water and purify it before the person drank it ... and maybe ... oh wait ... the only people that would need that are POOR and couldn't afford the $5000 let alone the $100 per month ... and of course there would be no profit in it so .... never mind. Hopefully I'll die before watching a robot do my dishes.
 
 
+1 Rank Up Rank Down
Apr 29, 2013
Robots have been envisioned for what...a hundred years already? To this day, there really isn't any useful robot out there that is accessible, affordable and does a useful task. It's in the same line as flying cars, good voice recognition and artificial intelligence...they all remain undelivered.

This doesn't mean they will not be delivered, they will. But people will worry about them once they finally arrive.

Also, and this is a more depressing take on it: we currently live in a world where middle class is under immense pressure, dissapearing even. The futuristic notion of the ultra modern home, full of immense screens and robots looks to become one of a richer upper class, not something for the average household to afford.
 
 
+5 Rank Up Rank Down
Apr 29, 2013
I'll jump on the bandwagon of pointing out that the Jetsons was made a very long time ago, and we're not much closer. I think the mechanics are the expensive part (as would be maintenance). To be able to perform a variety of tasks, you need a very complex machine. Unlike computers, machines are not getting cheap (think automobiles). And 3D printing won't cut it no matter how enamored you are with the technology. It will have some very interesting and valuable uses, but printing complex machines is not one of them. Joints, bearings, lenses, power sources, sensors, etc require complex materials and/or machining that will probably never be printable (unless/until we can print on an atomic scale). As examples, industrial robots are very expensive. A simple vending machine costs anywhere from a few thousand to $20,000 -- and they're not about to move around your kitchen and dining room handling china, glasses, and silverware.
 
 
Apr 29, 2013
Science has given us the hairless dog. If they can deliver a saliva-less dog, then it would get its exercise while you got your tennis balls fetched. Put some blinking lights on its collar if you had your heart set on a robot.

 
 
-2 Rank Up Rank Down
Apr 29, 2013
I honestly want to thank you for all your essays. I like their variety. Whether I agreed with them or not, they got my brain out of its decades long rut. In fact, they are more useful to me when I do not agree with your opinions, as then I am forced to figure specifically why I do not agree.

This led me to research, first at the library and then on the Internet. Most of the futurist books I had read were written by Arthur C. Clarke, Isaac Asimov, and Carl Sagan. Wonderful prescient writers, but becoming dated in light of new scientific findings. At this point I have found many good contemporary writers aware of present theories and technology. I am embarrassed to have written things, thought them pretty darned good at the time, and shortly later found others have said the same things years before, and a lot better. I wish I could say it is great minds running in the same channel …

For instance, Michio Kaku. Apparently he has been long on cable television and working as a science commentator on the networks. Your numerous essays about robots could be contained in his PHYSICS OF THE FUTURE long chapter about Artificial Intelligence. Kaku has given me a sense of optimism about the future, even though I will not see much of what he writes come to pass. The future is going to come about in its own way, no matter my opinion about it. I might as well enjoy speculating about it as I did when it stretched ahead much longer within my lifetime.
 
 
Apr 29, 2013
There are robotics competitions typically including university students that involve collecting balls or maneuvering balls into a goal area. I'm sure that could be adapted into something that will return balls to the pitching machine.

As for the rest, it's a long way off before machines can do complex household tasks more cheaply and more effectively than hiring another person.
 
 
Apr 29, 2013
I totally think you're right about this. I think we're a few short years away, and I even agree on the $5,000 price, and the reasonable set of robot chores.

I think the primary hurdles -- portable cpu / memory / storage, efficient general purpose operating system, computer vision APIs, practical limb design, decent batteries -- are all sitting on shelves waiting to be put into your domestic android. I think the overall software framework that integrates key technologies are next and I think they're far from impossible -- combining computer vision into environment modelling tools, object recognition, and creating kinematic tools. Proceduralizing tasks will be a cakewalk once that framework is in place.

I think the biggest impediments left are: strength -- servo-controlled fingers, shoulders, elbows, and knees are going to suck. We're not likely to fix that real fast. Robots will be pretty light-duty until we come up with better motion controllers. Sensors -- They'll need to touch things which will require an API about as complicated as computer vision. I don't think the sensors exist in a usable form and I'm sure the API doesn't exist. They will be limited without this. Batteries. More power = too heavy for servos to carry. Little power = limited range and frequent recharging. I don't think we're looking at robots practical for outdoor applications in the near future.
 
 
 
Get the new Dilbert app!
Old Dilbert Blog