Developing a robot worth buying

Or, you can take the Wal-Mart approach, and contract with a company that uses undocumented workers, an unknown number of which were not paid overtime or minimum wage. Lots of savings there!

Seriously, the straight-line cost is only a small part of the picture. A robot is a capital expense. In addition to the actual ticket price, a company must calculate its ROI, against using that money for something else. Money costs money, either in lost investment, finance charges, or whatever. A $400,000 robot actually costs more like $500,000 to $600,000. All this before service agreement, maintenance, upgrade, and operator's costs, of course. Robots don't just run themselves.

Hourly or salaried workers, OTOH, are paid incrementally, and costs associated with them are known. Once you've factored in direct employee expenses, there are no finance charges, maintenance fees, etc. for a human. In the end they're pretty cheap.

-- Gordon

Gordon McComb wrote in news: snipped-for-privacy@NOgmccombSPAM.com:

Yes, money costs a LOT of money.

Gordon , you're a man of vision. Robots that don't run themselves are going to be run and repaired by other robots.

Yep, humans can be pretty cheap, you get what you pay for. Robots don't (won't) steal. Or will they? They don't call in sick either. (Happened to me today, my chief tech called in sick, I had to go do his job, and of course I had a bad time of it. I feel as though today is the anniversary of Murphys law, and he is really celebrating.)

Joe

Was that meant to be sarcastic? I though that Gordon explained it fairly well for a single paragraph verion of expense economics.

I thought we were talking about things available now or in the near future. The $400,000 security guard was an *old* project.

Yes, sometime in the future we will develop robots that can repair other robots. Of course, these robots will also cost money.

Just imagine when a hacker gets into the robotic security system and empties your warehouse overnight. And of course, robots *never* break down. :)

-- D. Jay Newman

"D. Jay Newman" wrote in news:d98svd$oo0$1 @f04n12.cac.psu.edu:

picture.

No, not meant to be sarcastic. I've owned and operated a small business since 1987. One of these days, I'm going to actually make some money :)

(Happened

Hmmm, when I envision this scenario, I see the guardbot as a part of a complete security package. Fence, camera, moat with crocodiles, trained chimpanzees with AK47's, etc.

BTW, I am here to attempt to engage in constructive dialogue. If I start acting trollish, please let me know, sometimes I can be rather insipid.

Thanks,

Joe

The only beef I have is that you were posting some really good stuff to comp.robotics.misc and misc.business.product-dev (thus being a big help to me in trying to get enough critical mass to revive that dead newsgroup), then you went back to posting to comp.robotics.misc alone.

:(

"Guy Macon"

(Homer Simpson voice) D'OH!

Guy Macon wrote in news:11bgeo0sbqc25d3 @corp.supernews.com:

Sorry Guy, but I did not post to misc.business.product-dev. Ever. Someone must have dragged my posts along in a reply?

Joe

Hope I'm not too late for this discussion. I think one of the difficulties in discussing such a topic is that many of us are either in the same position that you are or have dreams to one day be. I have a few ideas that I think are great for example, but I selfishly don't want to share them.

I like robotics because the more i read about it and the more forums and chat groups i subscribe to the more I realize just how small of technical community it has and how much opportunity lies in its future. A few months ago there was an article I believe in Technology Review that said robotics today is where computers were in 1978. Also I'm an electronics nerd and nerds just can't help loving robots. They get us a lot of chicks.

I see two promising paths in robotics. One is to create mostly mechanical devices. Things such as the infamous Romba that require little intelligence and a lot of mechanics. A good starting market for these applications would be healthcare. Especially with the baby boom reaching old age (hope i didn't offend anyone) there will be a larger need for cheaper, easier, always available help, and healthcare has a lot of money. You can as stated before hire someone for $5.65 an hour to sweep streets or stock shelves, but 24 hour health service is a huge expense. Much of the cost could be avoided with proper monitoring devices and some assisting mechanisms. If your target is robotic hobbyists this may not be the path to take, since each application will require completely different mechanical parts. You could however market simple mechanical drivers or mechanical sets. Lego Mindstorms and the relatively new Vex Robotics

The other is to work on intelligence. If you transmit video and control of the bot wirelessly you can make a robot as intelligent as your computers and software allow at no additional power requirement. This is the route I'm taking. I think it'd be an attractive product of hobbyists. Just pop in your software and you've got a smart robot. Yet hobbyists, atleast ones in college like myself don't like paying for big software packages. I think this is a very promising area if you're looking long term. Intelligence is our limiting factor right now. We need a robot that can talk back to us, take complicated instructions, drive our car, clean our house, go to work for us. I think mechanically we can pretty much do that today but we're not even close on the software end. Who's going to design the os for robots? When windows won't even be used anymore because there are no PCs, just personal robots. If its me I'll give you each a free copy!

In my opinion, if the people working on the early computers in 1978 had selfishly not shared their ideas, it would have dleayed the revolution and given the glory to someone else who was willing to share.

Excellent point. I saw an ad yesterday about a device that lets an elderly woman move a quadraplegic spouse from bed to wheelchair without having to do any lifting. There is going to be a need for more devices like that.

I have become convinced that the hobby market is a losing proposition at this time. It really is oversaturated.

I agree. Putting the smarts in a standard PC with a radio or IR link is much better than using a microcontroller. It splits the task into two subtasks; [1] Can I make this thing smart using my 3gHz 500GB brain? [2] After I solve [1], how small can I make the brain?

This is not an area limited to or even the province of electronic nerds and/or software experts.

I don't think "big software packages" are the limiting factor. Creative new insights into what it means to be intelligent is the limiting factor. There are cheap and even free developmental tools. It is what you cannot buy, ideas, that is the limiting factor. Although if you lack the software/ hardware expertise you may be limited in your ability to implement those ideas.

This is essentially a newsgroup about hardware such as PID control and H-bridges etc.

Anyone who comes up with novel/new ideas about how to make a robot smarter will not be giving them away in a newsgroup :)

JC

If you are looking towards the full humanoid style robot I think we need to take a closer look at the way the control and intelligence functions are distributed throughout the body (its not all in the brain - apparently). This would indicate that we should consider implementing a hierarchical framework of controllers and computing units in some form of mesh connection.

The advantage of the meshed network is that, although more involved during the design and development phase, will end up being much more robust in operation. This also should help with making a more inteligent unit as the processing power required is shared out amongst a number of processors and thus simplifies the programming of the higher brain functions.

I am suspecting that the project, at the full humanoid android level, may take a significant amount of your remaining life unless you get a wider input from the right sort of willing people. Of course there are simpler tasks that can be accomplished without the need for an humanoid android format.

Actually, comp.robotics.misc is about any aspect of robotics and misc.business.product-dev is about any aspect of product development, so a discussion about someday being able to develop intelligent robots is on-topic in both groups.

I disagree. True, there will always be the paranoid inventor who thinks that his ideas are So Valuable That They Must Not Be Revealed, but for every person like that there is another person who realizes that having a small piece of a big pie is better than having 100% of a pie that never gets made. In addition, it assumes that the only way to protect an invention is with trade secret protection. Anyone who comes up with a novel/new idea about how to make a robot smarter

Marvin Minsky's Society of Mind theory hold that this is how our minds work now. See [

One could still do the development work on a standard PC by having the fast PC processor emulate a number of smaller processors in various mesh configurations. If total CPU power becomes the limiting factor, this would be an excellent candidate for a Beowulf cluster.

Robotics is essentially about hardware and its software support, not about AI or philosophizing about how one day robots might become intelligent. comp.ai.philosophy is a more suitable place for such day dreams.

I wasn't suggesting that you couldn't discuss about someday being able to develop intelligent robots, only that comp.robotics.misc has not been for that purpose.

If you come up with a novel/new idea to make a robot smarter you sell the patent to Microsoft, as they have the financial clout to enforce it, and retire to the Bahamas :)

If I could come up with an answer to your subject line, "Developing a robot worth buying", that wasn't already being done, I would not reveal it on the newsgroup.

There is plenty of useful stuff that can be exchanged that is of no personal commercial value. To give away a commercially viable good idea would be just plain neglectful to your family and yourself.

JC

Sorry Guy, I'll pay more attention to that in future (incidentally, I didn't do the setting of Followups but I will make sure to cancel them. Do you wish me to eliminate other ng's from the Groups list as well?). [%X]

Agreed. When I mentioned processors I was not distinguishing between hardware processors or software processors particularly. I tend to look at the overal architecture and structure then decide what has to be hardware and which parts software at a later stage of development than most others consider.

Incidentally, did you ever pick up on the Mentifex thread at all?

I rather like the crossposting. It brings in fresh ideas. I have set the automoderation software to allow crossposts to one other group, and that group can't be a moderated group for technical reasons. I also monitor the groups that are crossposted to to see if anuone objects.

I did, and am still thinking about it. It sure has spawened some interesting websites! Like these:

Robotics is essentially about hardware and its software support, not about AI or philosophizing about how one day robots might become intelligent. comp.ai.philosophy is a more suitable place for such day dreams.

I wasn't suggesting that you couldn't discuss about someday being able to develop intelligent robots, only that comp.robotics.misc has not been used for that purpose.

The average "robot" hasn't the brain capacity of a slug let alone any kind of potential for "human intelligence".

They are little more than animatrons which we can imagine are something more because they have a fancy PID system or can even balance on two wheels. This is what hardware/software engineers have been trained to do. AI is another ball game.

If anyone could come up with an answer to your subject line, "Developing a robot worth buying", that wasn't already being done, they would not reveal it on the newsgroup. They would be out there doing it.

JC

I agree, up to a point, but I also have a real-world counterexample; the AI used to animate the opponents in modern computer games is getting to be quite sophisticated. I see no reason why a robot can't attain the same level of intelligence.

So you keep saying. In the real world, things don't always work out that way. Look at aviation 1900-1920, for example. There was plenty of open discussion about aircraft design going on. Those with the answers to "how can man fly" did not always keep those answers secret. There were many people who had part of the answer or who had good ideas but lacked the resources to turn them into aircraft.