6 S's of Capable Home Robots

Gordon:

Most people have forgotten what the home brew computer market was like before the IBM PC made its debut. The main hobbyist period for personal computers was from early 1975 with the MITS Altair through mid-1981 with the IBM PC. During that time people would buy boards from places like Jade Electronics and Priority One Electronics, solder them together and build an S100 bus machine. The machines could not do much. I think that is the period of robotics that we are entering. The volumes and margins were low, and people saved money by building it themselves - Heathkit style.

Shortly after the IBM PC, the business model changed and volumes went way up and the hobbyist influence went way down.

By the way, the Readybot is designed to do both semi-automated tasks and be teleoperated. I would state that its goal is to be the S100 bus of hobbyist robotics.

Lastly, with inflation, the $2500 Mac 128K in 1984 would cost $4929 in 2007 (almost double.) A $5K platform in 2007 would have cost $1235 in 1975. Hobbyists were willing to spend a couple of thousand dollars on a personal computer (that did not do very much) back in 1975. BTW, I used the inflation calculator at:

to get these numbers.

I have no false expectations about what a robot like ReadyBot can do (not much), but it can be another mile post in the long road towards useful general purpose robots.

-Wayne

Hi all

I have to agree with Gordon on the cost point. This is a great robot but cost is a major factor in the market place. If you could earn money with one of these in a manufacturing environment say, it may be worth the money but as a home help it does look expensive. Another problem is the idea of tele-operation. If the robot was able to wash up on its own then this may be useful, but if it has to be told how to wash up, where each plate is. When to pick it up. When its clean. Were to put it to drain etc by a tele-operator then that lessens its value.

It looks like the same old problem robots have been coming up against for years. We need AI of a sufficient standard to use the hardware designs available to us to their fullest extent.

Until AI has reached this standard all the robots built for this sort of environment have to be focused on one "simple" task. ie: vacuuming or grass cutting. I think this is the one thing that is holding robots back.

I don't know if its possible to create an AI brain in the current computer hardware architectures we have or if we need a new type of computer to achieve this. I don't think anybody knows what AI really is. We know what we want it to do just not the methods to be able to do it.

I think a break though in AI will jump start the personal robotics industry but until then I feel it is a dream.

Just my personal opinions which looking at my current creations probably aren't worth a lot ;)

Rob

btw. When I stated the "simple" tasks don't get me wrong. I have not built a robot that could mow the lawn yet and would love to be able to do so, BUT i suspect washing up is FAR more difficult that grass cutting. I'm happy if mine stops before it hits the wall and drives in a straight line :)

See On Intelligence by Jeff Hawkins. Maybe this is the kick AI needs. Amazon listing:

-John O

Just checked it out on Amazon.

It made my head hurt :)

Its available second hand in the UK for about £7:00 including delivery.

Its on order as I type.

I suspect it will be above me but I'd love to understand a bit about how the brain works.

Weather it is possible to build a brain like system in silicon I don't know, but may be this book holds the answer.

Thanks for the pointer.

Rob

I was at a robot conference called Robobusiness and the AI question was presented to Keenan Wyrobek from Willow Garage

Keenan basically said that the experts believe AI doesn't exist yet for complex tasks so the initial focus of their PR1/2 robot would be on some form of teleoperation for manipulations. You can see some pretty cool videos at their website.

My view is that we need to get a capable/affordable platform in the hands of all the great programmers out there in order to advance the field. Imagine telling the contestants of the DARPA Grand Challenge that they have no access to robotic vehicles for development and testing. This surely would have slowed things down. In fact, I heard that the Stanford team that won basically got a robotic vehicle from VW so they could focus solely on the software side.

Danh

Actually, I've read the book and it is not too technical.

Actually, I've read the book and it is not too technical.

Agreed.

-Wayne

I briefly saw an article about someone building memory systems on this model. I'll see if I can find it next week, maybe there's an online copy.

-John O

If that's the case the period has lasted some 30-40 years. I'm not contesting the idea that we're still in an exploratory phase, but comparing this kind of robotics with PCs misses the mark, IMO.

The main reason for the disparity is that after the hobbyists, the early adopters of commercial PCs (IBM PC, Kaypro, etc.) MADE money on them. IOW, they spent money to make money. They used them in business, or even to start a business. Markets are different when a discretionary product is consumed for its own benefit only, which would be the case of a domestic robot. Except in very narrow cases, how can you justify its expense?

Because the market is not likely to be energenic unless the price is fairly low (under $1,000 or so retail) you just won't see many companies making the R&D investment. Without the investment we'll lag behind on things like sturdy platforms with arms and hands that can do at least what a toddler can do, affordable vision, reasonable safety (i.e. very strong but not too heavy), and other features needed to propel this forward.

Well, the PC legitimized it, but there were already lots of business users with even the Radio Shack TRS-80. That was the first computer I used to write articles with, and again, I was making money with it. There's no way I could have justified the expense for the sake of a toy, even to myself, let alone my wife.

Agree, and lots of us spend that much on robots today, small and large. Hobbyists have a different perspective than the mainstream market.

I'm just as much a fan of robotics as the next person here. But the problem I have making too large a leap into tech that doesn't yet exist, at a pricepoint anyone can afford: I've found it discourages people from getting in - and especially staying in - this field.

It's like research on nuclear fusion for power. We know how it works and what we need, but we're just not there yet in creating a containment vessle to tame the thing. As a result it largely remains a fantasy as a means to deliver electrical power. (The violent explosion part we have worked out pretty well, unfortunately.)

On the other hand, there are MANY robot markets that are workable today, with present technology, and at prices those markets are willing to pay. For example, security, law enforcement, and military are all willing to pay fairly high prices for a working telerobotic device. The US gobment is willing to put money into DARPA Grand Challenge because it sees good potential for autonomous vehicles (less robots and more like a car version of airplane drones, but we can call them 'bots just the same).

I'd be happy if I see a domestic robot take the market during my lifetime, but it might not happen. Lots of other robotics challenges will be met by then, and it's important to keep the potential for success in focus.

-- Gordon

Lots of people make the analogy. There is no particular reason why we have follow the exact same path and time-line of the PC. I just wanted to point out that the personal computer hobbyist era lasted ~6.5 years.

I would claim that there was plenty of product development going on. There were 100's of S100 boards developed, marketed and sold, for a profit.

People purchase things because they are fun. It is a hobby! How much money do people spend on hobby boats, cars, and even planes?

If you go back to old copies of Byte and Kilobyte magazine, it was littered with ad's for hobby grade products that people developed and sold for money. Today, flip through the pages of Servo and Nuts and Volts. Products are being developed and sold for a profit.

The reason why I chose the PC as the demarcation point, is because shortly after the PC was released, the hobbyist marked more or less imploded. After the dust settled, there was Apple and PC architecture. The rest were swept into the corner and forgotten -- Commadore, TRS-80, H89, Acorn, Pet, Cromenco, etc. all gone.

People were buying Apple-]['s to run Visi-Calc, the killer ap. of the PC industry. People ask what is the killer ap for robotics, and I do not know.

We are in total agreement here.

I agree. I have watched people decide to spend 100's of hours to develop a robot sub-system that is readily available for $10-$20. Sometimes I shake my head in wonder at how cheap amateur robotics people. Part of it is your fault for writing the book "Robot Builders Bonanza" and showing how much could be done on budget. ;-)

[snip fusion]

No argument.

I have no idea how long it will take. I am just stating that hobbyists have can continue to dabble in this area.

-Wayne

Musing...in the 50's when the modern idea of a human-like robot really took off, the futurists predicted all sorts of work-saving devices that would give us all the leisure time we wanted. Push a button and dinner comes out of a thing. Push another button and the house is cleaned. Etc. Some of that isn't too far from what I have now, but today as a society we're all getting weak and fat, and slowly realizing that the idea of sitting on our asses every non-working hour isn't so good for us. ;-)

But, the ide of the human-like home robot lives on. A guy I work with, who not coincidentally managed the development of the Heros, talks like Gordon...40 years now and we still don't know what they're going to do.

I want a robot that doesn't forget to take out the trash on Mondays. I want one that chops onions, garlic, and celery, but doesn't take 20 minutes to clean. Since I was a kid I dreamed of one that would stir the pot when making real pudding. :-) I want one that will play catch with me, and my son. In my house, I want HAL, not Rosie. I can get HAL...

:-)

-John O

--snip--

Wayne,

Assuming your last sentence was intended to read "how cheap amateur robotics people [are]":

I can only speak for myself with any certainty, but I'll guess that I'm not alone in seeing a difference between "acquiring capability" (through purchasing) and "gaining understanding" (from building something and trying -- and trying and trying -- to make it work). I don't see these as exclusive; in fact, I suspect that anyone who spent hundreds of hours making his own IR-based wheel encoders is more than a little likely to spend $20 for "off the rack" the next time he wants one.

Ah, well, I have to agree with you there. Any justification I feel I need for my electronics and construction budget will begin with the phrase "Gordon made me do it!"

Frank McKenney

-- It does not do to leave a live dragon out of your calculations. -- J. R. R. Tolkien

-- Frank McKenney, McKenney Associates Richmond, Virginia / (804) 320-4887 Munged E-mail: frank uscore mckenney ayut mined spring dawt cahm (y'all)

The point I was trying to make, and obviously I missed the mark, is it is not necessarily the price that turns people off but the usefulness. We need to worry more about what it does than what it costs. Cost must be kept in mind but function has a lot more to do with it.

Then we need to consider marketing!. The Amiga computer ran circles around both the Mac and the PC when it came out but a lack of support and ZERO marketing killed it. The computer was a great tool and was followed by many but good old Commodore shot it in the back due to poor support and total lack of marketing.

I believe the key to success is finding the RIGHT NICHE and fill it with a quality product even if it costs a bit more to start. At this point trying to duplicate "Data" of Star Trek is not a realistic target but what about something simpler but functional. Maybe a home bar tender, a intelligent "stand up" wheel chair, who knows....

FWIW a MAC II when it came out] I believe listed for like $4999 and a IBM PC was a touch more with it's whopping 5 meg hard-drive :) Then the Taiwanese got involved :).

Monty

Yes, I know, and the comparison has never proved itself out. *I* even used the comparison in the mid 1980s when I was pitching my original book, but I was wrong. After several decades at least some of the predictions should have come true by now.

What we have are two very distinct genres, the hobbyist market that tinkers with robotics because (as you say) it's fun, and the BUSINESS of creating a home robot. Danh was saying we have all the technologies now, which to a point is true, but not at prices anyone but the very well heeled can afford.

To get the prices down we need scale, as was done with PCs. But robotics is vastly different than PCs: Scale in robotics is FAR more difficult to achieve than board-level electronics, and Moore's Law simply does not apply. We can't expect manufacturers to create precision metal gear motors of the type needed for higher-end robots for a buck a piece. Motors are made in high quantity now for all types of industry, and they continue to be expensive. Why? The parts that go into them are expensive, and constructing them takes more than a fab line at a board sweatshop. We know a lot of the cost (and weight) of a our hobby robots go into motors, and many of us use motors that they made in large quantity.

There will simply never be a market for domestic robots unless the price gets to a point where its value encourages interest. The value of a domestic robot is in direct proportion to the capabilities afforded by its construction. The type of construction needed to produce a robot that can actually do useful work (and hence "earn its keep") is significant.

So what we end up with is a zero sum game: we can't have low-cost domestic robots until they are cheap and do useful things, and the only way they can become cheap is large scale, which won't happen because apart from a few vacuum robots they just don't do useful work.

Contrast with PCs: even the early computers with 64K RAM and a calculator chip for a brain could do Visicalc, a "killer app." For the PC, the killer app didn't take a lot of hardware. This is why the PC-robot comparison falls flat.

So, says I, it's better to take all this effort and build a robot that SOMEONE WILL ACTUALLY BUY! I've already talked about the high end, but on the low end there's still the hobby market, the toy market, the technophile market (try in 3-4 years because of the economy), and the growing service market for specialized mid-level bots for things like well inspection, pest and vermin control, and so on. These are (or can be) just as exciting as a domestic bot, yet the probability of someone actually managing to build a successful business with one is far greater.

Success even in a hobby keeps you interested; decades of false promises just make you look for another career path. This is why I'm not keen on spending lots of time on a so-called domestic robot as I think the whole thing becomes a kill joy. ("Gee honey, but look, it fetches me a beer so I don't have to get up out of my chair!" Yeah, right!)

-- Gordon

I wanted to address this part separately as I think this paragraph misses a critical point. The idea of a PC is that an ordinary person can adopt it to his or her work - *anyone* can make money with it, not just someone selling back to the same small audience they bought from in the first place. The ability to justify the acquisition makes even thousands of dollars palatable. Hobbyists don't need the same kind of justification. They spend money for other reasons.

The vast majority of people who ended up buying a PC never read Byte, 80 Micro, or any of the other magazines of the time, or walked into a Byte Shop. A truly functional domestic robot is aimed at the same user. That user is expecting a qualitative cost-value ratio, and that ratio is highly unlikely to occur in the way PCs translated from hobbyist trinkets to mainstream markets. (Which is why, after decades, it still hasn't happened.)

In closing, I liken this whole discussion to building a rocket to send humans to Saturn just because the Wright Brothers demonstrated their airplane at Kitty Hawk. Saturn is technically possible but not realistically feasible at this time. However, that shouldn't stop people from developing what they can now, proving the technology, and moving forward one logical step at a time.

-- Gordon

Gordon:

I think we are talking to cross purposes here. I keep talking about the hobbyist market and you keep talking about the high volume business market.

As far as I am concerned, the hobbyist robotic market is alive and well. People are designing, marketing and selling products to that space. I have no idea if these products are profitable, but they are unlikely to be able to pay for advertising if there is not some positive revenue coming in. I could care less if they are high volume.

People are already selling hobbyist robotic platforms in the $500-$2K range. RoboOne robots are a perfect example.

My statement is that if somebody designs and markets a domestic robotic platform for $500-$2K, robotics hobbyists will buy them. Period.

The ReadyBot domestic robot platform can easily be redesigned to sell in the $1K-$2K price point. When I did a marketing study at the last HBRC meeting, a few hands went up at $2K and a dozen or so hands went up at $1K.

-Wayne

[snip rocket analogy]

Gordon:

Again, I am talking about hobbyists and you are talking about post hobbyists. Until people developed word processing and spread sheet software, the original PC's were basically useless. None-the-less, hobbyists continued to purchase them. There was a thriving market for S100 boards to build out these hobbyist machines.

-Wayne

I would buy a capable robot for $2,000 if it had useful arms (Readybot is the closest thing). In contrast, if the Whitebox Robotics robot were $2,000 instead of $6,000 I still wouldn't buy it because it's nothing more than a PC on wheels with a webcam. It's like trying to sell the benefits of a maid service for the home but all the maids are missing arms!

Danh

Danh

Personal robots are still not happening, but great progress is being made on killbots.

I recommend reading "Wired for War, The Robotics Revolution and Conflict in the 21st Century", by P. W. Singer. This will give a sense of just how far military robotics has come in the last few years. If you haven't been following this, you may have no idea how many thousands of military robots have already been deployed.

For a sense of what's coming in the near future, search for "Legged Squad Support System". This is the production version of Big Dog. It's not even an R&D program; it's out of DARPA's Tactical Technology Program, which means it's one step from being fielded and used in combat.

John Nagle

We are and we aren't. What I'm trying to say is that the domestic robot that people envision requires a hefty R&D allocation that - frankly - even a group of dedicated hobbyists can't readily overcome. Been there, done that. The concept of a domestic robot is hardly new, yet for over

30 years we've barely made a dent in progress. There are reasons for this.

What we do have NOW is a great learning tool, and the hobbyist market (alive but not as well as it was in 2000) certainly exists. What I'd like to see is more development of the things that make practical sense, that propel us one step at a time. People are still running line following contests because it has a meaningful goal.

I'm not saying we should all stagnate just with the 10-15 year old challenges, but we're not even mounting a good defense against those - many robots fail a line following test as the line gets more complicated, or travel gets faster. I only shudder thinking of the 100X or 1000X complications of what a domestic robot needs to do. (And like I said, when people don't have successes, they tend to just find another area of interest.)

I'd say this isn't a hobbyist product as much as an educational product. In fact, at least in my experience, what you call the hobbyist market is really education. I'd say 80%+ of the sales I get at Budget Robotics are to schools and students. They are "hobbyists" in the general sense, but they aren't just playing. They hope the skill and knowledge they acquire with platforms like RoboOne will someday land them a good job in robotics or an allied field.

Because they are investing in their education the budget adjusts for it. I don't think they'd sell the RoboOne's they do just to guys and gals at home with too much money on their hands.

Your user group is a little skewed as a representative segment of the market. I know the one we had in San Diego, while it lasted, was the same way. Funnily enough, it was the students that sometimes came in that didn't stay around. They didn't have the time!

-- Gordon