What does 'robot' mean to you

[HarryBotter]

You're invited to provide your definition of what 'robot' means to you in 100 words or less.This contest was spawned by the following thread.

http://www.botmag.com/forum/showthread.php?t=240

Judging will take place on November 15, 2006 and the winning definition will be featured in the upcoming issue of Robot magazine.

The winning submission(s) as determined by the editors will receive a one year subscription or extension to Robot Magazine.

[HarryBotter]

I can't believe that there isn't someone out there who thinks about what robots are and wants to share those thoughts with others.

Remember there are no wrong answers here, just insights into what robots mean to you. And don't be scared by the 100 word limit, 10 words might do just as well.

Harry

[PeterD]

Ok, I'll bite. Maybe these ideas will stir some converation.


A robot is a machine that operates under self-control, performs a task, and reacts to its environment.


By my definition an R/C vehicle would not be a robot [no self-control, no reaction to the environment], but the Mars Rover would be.

What do you think ?

[dporter]

Definition:
A robot is a tool that incorporates and responds to information.

Explanation:

·A tool:a device to improve work

·Work:moving a force over a distance

·Incorporates and responds to information.
What makes a robot different from other tools (hammer, club) is that it can respond to information. It also has some information (logic) programmed into it as part of its design. A light-tracking robot is an example. The mechanical design incorporates logic -- one side or the other detects light, causing the motor on the opposite side to move. The robot responds turning towards light

[PeterD]

Definition:
A robot is a tool that incorporates and responds to information.


Would a BoeBot meet this definition?

Is is a device that improves work?

[dporter]

I think it would fit. When a boe-bot moves with a force over a distance then it is doing work. Improves, just means as a machine it increases force (more than what you put into it activating it) or increases distance (more than what you put into it activating it) on whatever it is you want to move over a distance. In this case you want the robot to move a distance or do something without you pushing or moving it yourself.

(Work input = Force input x distance input) = (Work output = Force output x distance output)

Work input = what you do to the tool/robot Work output = what the tool/robot does.

TGood question, thanks for helping me clear up what I was trying to say.

[dporter]

Peter, would the light tracking robot I used as an example be under self control by your definition?

"A robot is a machine that operates under self-control, performs a task, and reacts to its environment."


It would only be responding to the presence or absence of light, it couldn't control the light or how it responds to it. I have heard some people say that the light tracker is not a true robot, would you agree?

Thanks

Doug

[PeterD]

Hey Doug,

A light tracker _is_ a robot by my definition. It meets the three tests I laid out.

#1) it operates under self-control -> it makes its own decisions; there is no external operator making them.

#2) it performs a task -> avoids or moves towards light

#3) it reacts to its environment -> changes in the environment (presence or absence of light) cause it to act.

[Gary]

It's really quite difficult to sum it up in a concise definition - but how about:

An adaptive device that can interact and respond phsyically to it's environment, and perform a task - and/or that allows a remote operator to do so, either in real time or via preprogrammed command.

- the key word being adaptive.

EDIT: - last phrase added to clarify.

[dporter]

Hi Gary,

Thanks for moving things on with your definition. So from your definition it seems that you would consider a R/C car a robot? Also by adaptive I am taking the idea that you mean it can be changed, improved or modified based on conditions the robot may encounter?

Thanks, Doug

[Gary]

It's very tricky, skirting around this idea of whether to include devices without any autonomous capability whatever.

The problem is by requiring some degree of automony, you rule out things that are very much robotic. A dedicated teleoperated robot arm in a surgical theatre, for instance - has [for obvious reasons] no autonomous capability, and is directed by a skilled operator. I wished to include devices like in my definition - there is no easily definable boundary between such a device, and a simple R/C car, and you have to dive into the muddied waters of semantics, and try to define based on what the device can achieve, and how it does so.

In answer - a R/C car is not a robot, my definition requires it to allow a remote operator to perform a task, which a simple vehicle cannot [unless we get into semantics and lower our expectations of what constitutes a 'task']

[perhaps meaningful task would be more apt.. but still the same problem with what exactly that means, I suppose!]

If you add a manipulator to a R/C chassis, though, now you are arguably on the robotic bell-curve, in my opinion - and if you then add a camera, then definitely - go and disarm some bombs with it, and you're acheiving a very meaningful task! The same device could conceivably allow you to go milk a particularly dangerous cow, or any amount of other tasks which demonstrate it's adaptability.

So, by adaptive, I mean in the sense of what tasks the device can successfully achieve - by whatever means. So that includes either by remote control, preprogramming, or real time self-modifying behaviour - the important thing is that, ultimately - it can do it.
A R/C car is not adaptive, it can move around, push things, run things over etc. - the subset of tasks directly achievable by these methods is not sufficient to satisfy the definition.

- 'What is a Robot' - A deceptively simple question!

-------

A slight variation on this - do people consider a Robotic device - a Robot ?

Pondering this - I think my definition is for a robotic device, [i]not a Robot.
In this case, I think the two classifications probably are separated by possessing autonomy, or lack thereof.

[dporter]

Very well said Gary. I agree with you. I wouldn't consider an R/C car a robot either. I also agree that there is no clear cut line to where a R/C device becomes a robot. I like the ideas of a "Task" and a difference between a robotic device and a robot as maybe being different things too. So for others following this thread should the two be defined differently?


Thanks
Doug

[TheDuck]

Wow, some great discussion. I like that we all realize this is pretty subjective. The arguments presented definitely don't draw a line and it's hard to do so. For example, a sensor-light above my garage might be a robot in that it responds to its environment, adapting to the amount of light outside, without input from me, providing a useful service.
The catch is there's a lot of subjectivity to deal with here. I think adaptability is difficult to pin down as that it just a response to inputs as well. Useful is also subjective in that what I may find useful, you may not. For example, a robot that ties my shoes seems silly but what if I had no arms?

So I still don't have a concrete definition of a robot but what I like to see in my robot designs is:
- feedback on outputs -- don't assume the wheel turned for 2 seconds, KNOW it did and react accordingly
- another way to say that point is a robot that alters its state to match a condition -- instead of turn 90 degrees and hope I'm pointing the right way, rather keep turning until I am.
- have some sort of objective and be able to measure progress or effeciency against it (another example of feedback) -- I don't think we're ready for all-purpose robots yet and trying to make robots that are like people seems redundant since we already know how to make people :)
- I think redundancy is often overlooked -- swarms and even redundant systems, sensors and actuators on a robot seems far more robust

Or we could just check dictionary.com. http://dictionary.reference.com/browse/robot

The definition covers everything from simple machine to exact duplicate of a person! ;)

[Gary]

an interesting thought experiment you might want to try is to imagine something that perfectly embodies your personal idea of a 'Robot'

- now, start removing things from it in turn - and at each stage, ask yourself whether you still consider it a robot. Try it with the physical things, then the electronics things etc.

It's quite effective at helping to disseminate your personal definition.

[ben-ja-men]

A robot is the physical embodiment of a computer, capable of performing symbol manipulation of one physical phenomenon into another, based on its internal program and the information it receives from the external world. Much like John Searle's man in the Chinese room a robot has no meaningful understanding of the tasks that it performs.

[ben-ja-men]

Gary i believe the boundary between a robot that are not autonomous and a R/C car is that the car itself provides no feedback to the operator (apart from its observable movement in the real world which is more an effect not feedback as there is nothing inplace to measure or quantify it), essentially the feedback from the robot embeds the user into the control loop of the system making it autonomous.

the feedback recieved from the teleoperated robot is such that the user can continue input information for the correct operation of the robot without seeing the robot itself as such the person acts as the autonomous agent in the system. In the case of a R/C once the effects of operation are no longer visible to the operator the R/C is no longer able to be correctly operated as the entire system is disconnected. In much the same way that a skill tester machine is not a robot as if you couldnt see the effects of the inputs the system provides no feedback to the operator.

[dhylands]

I have a definition posted on my website (top of this page: http://www.davehylands/Robotics)

Robotics is the art of syncopating the arts of electronics, mechanics, structural and systems engineering along with a healthy dose of creativity, curiosity, aptitude and spare cash into a cryptic device that digests immense amounts of information about its surroundings, then discards it and drives into walls.

This original came from somebody with a handle of: Triffid Hunter
I can dig up an email contact for him, if required.

[HarryBotter]

Please keep posts to this thread limited to contest entries.

Any discussion about entries or follow-ups can be directed to:

http://www.botmag.com/forum/showthread.php?t=240

Thanks, Harry

[tatwood]

The contest for the best description of a robot is just the beginning of what may be a raging semantic debate among friends and enthusiasts for some time to come. We have picked out winners for the initial phase of this fascinating discussion, and like the home robot photo contest, will run this contest again for three months with the next round of winners to be announced February 15, 2007.

The robot definition contest may find intellectual breakthroughs, splits in the road, and even end points that become the “status quo” until the next insight has been published and accepted. This contest is philosophy, not just snapshots in time. The winners of our first robot definition contest are a few thoughtful entrants:

Peter DelMastro proposed:

“A robot is a machine that operates under self-control, performs a task, and reacts to its environment.”

Peter wins a one year subscription for this trenchant stab at a concise definition.

Doug Porter offered this refinement:

“A robot is a tool that incorporates and responds to information.”

Doug’s definition impliedly incorporates the concept of work by using the term ‘tool,’ and interestingly, brings in the concept of responding to informational input, which suggests a more powerful interplay than just “reacting” to the environment.

Gary then added the following element of teleoperation:

“A robot is an adaptive device that can interact and respond phsyically to it's environment, and perform a task - and/or that allows a remote operator to do so, either in real time or via preprogrammed command.”

Gary subsequently added that there are two concepts here: “I think the two classifications probably are separated by possessing autonomy, or lack thereof.”

Robot awards Doug, Peter and Gary one year subscription extensions for their thoughtful foray into the definition of a robot. We’d like to add MythBuster Jamie Hyneman’s comments on robots to the discussion—he noted that exponential in an RC transmitter that moderates human input and smoothes-out shakiness is indeed a robotic input, and we agree.

Having said that, who will meld these definitions into one overarching paragraph that pulls all of these elements together? We will find out on February 15, when the next iteration of our home robot photo project and our next robot definition contest will conclude, and at which time the next round of winners will be announced! The lucky winners will receive not just a year’s subscription to Robot but also a lasting presence in our contest archives. Our storage media will be updated as technology evolves, and the history of these contests will endure.

In the meantime, jump in and give us your thoughts--join the fun!

Tom Atwood, Editor-in-Chief
Toma@botmag.com

[PeterD]

This topic has generated a lot of thoughtful (and thought provoking!) ideas. It has been fun to participate in the conversations.

I'm flattered to have my ideas recognized.

Thank you!

[wrighthobbies]

Ok, I'll bite. Maybe these ideas will stir some converation.


A robot is a machine that operates under self-control, performs a task, and reacts to its environment.


By my definition an R/C vehicle would not be a robot [no self-control, no reaction to the environment], but the Mars Rover would be.

What do you think ?

Hi Peter,

This is a tough one since computers and embedded systems muddy the water quite a bit. Here's a twist for you - I can think of a machine that operates autonomously, but also takes direction from people. It reacts to it's environment and even causes change to its environment. It can track events and alert people when something needs to be done. It can alter its behavior based on time.

What is it? An electronic thermostat...

I would change your description a little to "A robot is an animated machine that operates..."

This is an interesting discussion!

[PeterD]

Hi Eddy,

I've thought about this idea on and off since I posted it. I am not sure if my definition is too broad or if my preconceptions about "robot" are too narrow.

For example, a vending machine could be considered a "robot" by the definition I proposed. That doesn't feel right.

On the other hand, we can probably agree that a Roomba is a robot. It looks & "feels" like a robot. However I think it has less functionality & complexity than a vending machine!

Obviously a Roomba is mobile and a vending machine is not. But I don't think that mobility is a criteria for "robot".

Maybe there is another aspect to this which just isn't clear to me yet.

[wrighthobbies]

Hi Eddy,

I've thought about this idea on and off since I posted it. I am not sure if my definition is too broad or if my preconceptions about "robot" are too narrow.

For example, a vending machine could be considered a "robot" by the definition I proposed. That doesn't feel right.

On the other hand, we can probably agree that a Roomba is a robot. It looks & "feels" like a robot. However I think it has less functionality & complexity than a vending machine!

Obviously a Roomba is mobile and a vending machine is not. But I don't think that mobility is a criteria for "robot".

Maybe there is another aspect to this which just isn't clear to me yet.

Vending machines, hmm. I soda vending machine may not feel like a robot but what about one that serves cooked fries? It takes frozen fries, flash-fries them with hot air, puts them into a cardboard holder and serves them to you. Would that feel like a robot? We embed automation and intelligence into so many machines today, it's getting harder to distinguish.

[Kevin Berry]

Well I've now seen this exact discussion, twice on two different forums, with very intelligent people, arrive at the same conclusion. For example, a modern fighter missile meets virtually anyone's definition of a robot. And maybe it is.

I usually back into this thought train from the other end. Being a major classic sci-fi buff, I know the originators (Asimov, Heinlein, and all those other pulp mag started geniuses) worked from a HUMAN direction down, not a MACHINE direction up. A robot replaced, relieved, or improved on human performed functions. So to me, and this is more a cultural definition than an engineering one, a robot is a device that mimics human behaviour, is task oriented, and exhibits some form of judgement while performing the task.

Applying this to real world, I recognize that this definition would include the french fry machine, as well as many software-only "bots". Just like humans, some robots are pretty dumb with limited judgement and some are very very smart but have no common sense!

[dporter]

A robot replaced, relieved, or improved on human performed functions. So to me, and this is more a cultural definition than an engineering one, a robot is a device that mimics human behaviour, is task oriented, and exhibits some form of judgement while performing the task.



The original idea of a robot was to mimic human behavior. What has evolved though are robots that take the best of many biological systems. Spiderbots that crawl about, others that swim, fly and track the sun like sunflowers. So maybe we can reword your definition to add that robots mimic living systems.

Doug

[Kevin Berry]

Doug, I like your distinction. As I wrote that post, I had put in a mention of how a robot definition would look a lot like a chimp or trained dog, but took that out. Now it needs to go back in. So we're heading down the road of a robot being a manufactured item that mimics a living function ...

but I still think pseudo judgement like qualities matter to the perception. An artificial heart? An ATM? Course an ATM "judges" if I've fed it counterfeit money or not.

Well I've managed to confuse myself!

[dporter]

Kevin, I think you're on track with the idea that there should be some way to incorporate judgement. In my definition I said that there is a "response to information", but making judgements or decisions seems more accurate. I don't know the specs on an artificial heart, but if it senses changes in oxygen or CO2 levels and can adjust the heart rate based on those measurements then I could see including it with robots in some way. That would bring us back to the discussion earlier in the thread on if a robot and robotic device are the same.

Doug

[JonHylands]

Personally, I think a device has to be mobile, and capable of a degree of autonomy, in order to be classified a robot. Yeah, that precludes "robot" arms and true ROVs, but I don't consider either of those to be robots.

A degree of autonomy means the robot can sense aspects of its environment, and react to those aspects. Having a cruise control on a remote-controlled vehicle would not classify it as a robot, but having a system that stopped the vehicle when it sensed an obstacle ahead of it would.

Basically, it comes down to having on-board smarts. That could involve something as simple as a photovore, or as complex as a DARPA Grand Challenge vehicle.

And yes, under this definition, ASIMO would definitely classify, because it has on-board sensors for balance, even though it is mainly remote controlled.

- Jon

[dporter]

Sounds like our split in robotic devices vs robots may be based on autonomous motion then. With the arm, heart, and rc car falling in the robotic device area. These devices have robotic qualities, but would not be true robots.

[PeterD]

On Wikipedia (http://en.wikipedia.org/wiki/Robot), the definition for robot begins:

"A robot is an electro-mechanical/bio-mechanical device(s) that can perform autonomous or preprogrammed tasks."

I think that autonomous activity is a key indicator of a robot. I don't mean the ability to walk, fly or drive around ... I mean actions that are autonomous of human direction.

The robotic arm in the machines that I am most familiar with does much more than replay pre-programmed sequences. For example, it has the capability to, on command, calibrate itself to its environment. By that I mean: it measures the interfaces at each of its workstations and then adjusts its teach points & trajectories based on what it detects.

This is sophisticated & complex functionality that allows this robot to adapt to its environment.

A further enhancement, or evolution up the robotic scale, would be to have the machine automatically detect the need for self-calibration and then perform the job without being told. Kind of like having your kids realize their room is a mess and clean it up on their own. Wow, that would be great! :-)

[TheDuck]

You'll have more success designing a robot to do it than getting your kids to do it. lol

Actually, something I'm always saying is lacking in robots is feedback. When I was at the Firefighting robot convention I noticed that no robots were aware of their own actions. For example, they assumed they turned 90 degrees to find it was 91. After a couple turns they were thudding into walls. Then they found some heat source ahead of them and dispensed a fire retardant in roughly that direction and missed it half the time. So the sensors produced something to do but weren't used to confirm successful execution of the subgoal. Of course, having said that, I see more and more circuits that register a resistance to things like motors and servos moving and send that back to the IC or control system running the show which is good.

So I think feedback is important to successful execution (I'll stop short of calling it "self-aware" LOL). Any comments?

[HarryBotter]

I was just starting Bekey's book "Autonomous Robots" and stumbled on his definition of robot, a machine that senses, thinks and acts. I think I like it although it lacks full clarity in that the definition of "thinks" is wide open.

It requires some level of perceiving its environment, a processing of that information and then an appropriate reaction.

I apologize if we've been close to this before, I just haven't read all the posts in this thread. :)

Harry

[wrighthobbies]

I like "perceives, processes and acts" better than "senses, thinks and acts". But this goes back to my earlier example of an electronic thermostat. It technically fits this description but wouldn't be considered a robot by most people.

[jabstarr]

this is indeed an interesting thread!

i think it is essential to realize when discussing robotics that the technology is developing very rapidly at this time. cheaper development tools, which are more easily accessible are expanding the number of potential developers. you do not need to be an engineer to create mechatronic systems anymore: designers, artists and other formerly unknown actors are joining in offering a radical innovation, challenging the conception of robotics itself. new products are emerging as the line between mechatronics, robotics and interaction design is being blurred.

in other words: massive changes are happening fast.

and i think it is important to differentiate between what robots are/use to be and what they are becoming.

for example the osaka international design competition 2006 (http://www.jdf.or.jp/robot/en/index.html), had the theme ROBOT in a more widely defined sense:

“The term “ROBOT” here means not only humanoid robots, but also any other tool created to solve present-day social challenges.

We define “ROBOT” as new interfaces that can reduce people’s loads and burdens in their daily lives and that help establish more delightful and positive relationships between humans and things. We look forward to design proposals with the concepts and techniques of ergonomic design and universal design incorporated, and we hope these proposals can help make our lives safer and more comfortable, rather than striving for material wealth alone. “

and the winner was the beautiful dandella project:

http://newshub.nus.edu.sg/headlines/0702/images/dic_2.jpg

“A hand-held Global Positioning System (GPS) tracking device, Dandella helps guide those who are travelling or have lost their way by always physically pointing towards their intended destination. Combining GPS and robotic technology, Dandella units can also be programmed to track each other and their buds (pointers with light pulses) respond by pointing towards one another. This intuitive design feature also allow users to locate each other by just following the direction which the bud points – making Dandella useful for both the young and old.”

- - from the national university of singapore homepage, you can check out the dandella project here (http://moduleblog.nus.edu.sg/blogs/id4104/archive/2007/02/13/nus-design-incubation-centre-dic-won-gold-prize-from-international-design-competition-osaka-2006.aspx)

does the dandella project qualify as a robot? or does this represent a new approach to robotics?

i know i'm well over the 100 words now. i get carried away ;) but since the competition is over, and i think this subject is far from dead, i hope you will bear with me. so i'm posting these considerations, hoping to hear your thoughts on this.

are we on the verge of a paradigm shift in robotics? is it time for a new scope on robotics? perhaps even a new definition?

sincerely,

jacob remin

[tatwood]

Hi Jabstarr,

Thanks for the thought-provoking comments and the great pictures. My view, as a student of evolution and as an observer, in the position of an editor, is that the flowering of robotics will insinuate itself into all aspects of our lives in ways that can be projected as science fiction now but that will be discovered, in some respects, later as our technology continues to unfold, and only then will be characterizable in handy, practical and philosophically tidy ways.

My personal hunch is that significant paradigm shifts will be part of the process. It is a multivariate thing happening on so many levels--robotics have been identified as a learning accelerant in education and may be a key catalyst for the successful education of next generations in more than a few societies. Those countries that successfully inplement robotics curriuculum and programs to bootstrap their technological infrastructure forward in an ever-more competitive world may be using automobiles that park themselves and faucets that turn themselvs off with most people never connecting the dots.

Clearly, at Robot Magazine, we are pro-robotics for reasons that go far beyond publishing an entertaining and hopefully educational hobby publication. There are huge forces and undercurrents at work that will ultimately benefit all of us as technologies converge. Just my two cents.

Tom Atwood
Editor in Chief, Robot
toma@botmag.com

[Kevin Berry]

"new interfaces that can reduce people’s loads and burdens in their daily lives and that help establish more delightful and positive relationships between humans and things"

OK I really really like this - robots are a tool that (actively?) (without full time human supervision?) assist humans in managing objects (and, to be fair, managing information, pulling in the software definition of a "bot").

Thank you for a most interesting post.

[jabstarr]

hi again,

just thought i'd let you all know, that my bachelor thesis "a new approach to domestic robots" is done :D within it is an extensive discussion on the definition of the word robot.

the whole report can be downloaded via this link (http://campingsex.org/bachelor/media/BACHELOR_lores.pdf) (25MB .pdf)
for a brief introduction to my ideas and thoughts, take a look at this post (http://www.botmag.com/forum/showthread.php?t=479)

any kind of feedback is much appreciated!

best regards,

jacob

[josephco_3637]

cool,.,. a robotics student,.,amazing


i think robots are programmed computers that has a wide range of capabilities such as doing a simple or complex task physically