Let's Make Robots!

Only autonmous robots are real robots!

Whatching TV, some guy displayed some "installation art", with a "robot horse", some mechanics controlled by electronics, looking oh so funny.

"That´s not a robot", I said.

"Why is noone elses robots but yours entitled to be called a real robot", my wife said.

I tried to explain to her, that what we just saw on TV was an installation, not an autonomous robot!

It is jus tlike "Robot Wars".. These things are NOT "robots" in my terms. They are "Remote controlled things"!

The "robot-magic" happens when you let some logic control mechanics, based on input, and the mechanics change the environment that again resoults in new, not pre-programmed inputs to the logic, that then..

Then you have robot-life!

Then you have an autonomous robot.

Then you have a real robot

(in my terms :)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

So I have built a thing that has motors, a microcontroller, a speech recognition board, some servos... I turn on the thing, say Forward, it goes forward, I say Stop, it stops, I say Pick-Up, it picks up an object, I say Turn, it turns, I say Stop, it stops, then Forward, Stop, Drop... And here is my thing that brings me an object (a beer?) obeying my voice commands. Is that thing a robot? OR is just a remote controlled (by voice) vehicle?

Consider the fact that  language is like the face of a cliff. Every once in a while, a storm comes along and erodes the cliff away, until one day, there is no sign of the cliff.

Everybody has a different definition of what a robot is, few are based on any sense of reality. Consider the fact that Rossum's robots were biological, not electromechanical. This makes every robot we've ever seen something else.

The technical world is full of examples of words that have been bent out of shape so far that only a few old timers even remember their original meanings.

I am a hacker. Always have been. Over the last thirty years, the media has taken the liberty to redefine the word to the point where once it was a badge of honor, now it just means criminal. So it is with robots. A robot is a robot by virtue of being called one, since there are no 'real' robots out there that we know of and the non technical world has long ago redefined the term for us.

I would maintain that the distinction that makes a robot a robot is the ability to make decisions based on ir's own sensory input and internal memory without the interaction of a human. A remote control vehicle requires human input to do anything and has no self deterministic capabilities. Having said that, almost all robots start out as a remote control vehicle while the builders learn enough to give it some level of autonomy. That technically is as little as the ability to avoid obstacles. Considering that almost all robots retain their remote control primitives, I would be inclined to grant the status to anything that has even the simplest processor 

I think that the most basic of electrical and mechanical systems mixed togehter, aka a motor and battery is not a robot. It must have some underlying logic or function, eg a motor and battery doesn't. BEAM bots are about as low as you can go in my world and still be considered a robot.

Tinhead because it is a useful tool to build robots. It was not a critique of any projects on Lmr. There is a lot of skill and talent put into projects like valkyrie and you have my respect for what you have achieved.
... shall I continue updating the Valkyrie page or not? I guess it's no robot by this definition ....
This debate is not about our robot pages on LMR, it's more about the man and woman on a couch commenting on the world around them. And I'm glad he shared it with us.

Thanks for starting this thread. Firstly I'd say I agree almost 100 percent with you Fritzl. There are a lot of things passed off as robots which degrade the word Robot. Remote control is not autonomous and should only be called remote control not robot.

I think  multiple sensors and sensor fusion is essential to give awareness to the robot and to equip it better against the uncertainties in the world and to eliminate the need for help from a human. If it succeeds at this then it is a useful robot.

In defence of coffee machines though, I'd say I'd call it a robot if it could ask me how I like my coffee,understand what I mean by it and could produce it exactly the way I like it and maybe even give me some advice about the weather today before I left home.

But I guess that is taking us into another area that is still a mystery to the best. What is consciousness and how do we make artificial general intelligence ?


Your explanations tell us about your passion for robotics.

From your explanations I get the feeling that If I want to make the robot I must be the robot.

Well done and thank you

Here is the truth! (No humble stuff here)

"Does a robot need to have more than one source of input to be a robot?" Yes. Unless this one source is mixed. But only one allways makes it very simple. Very. Dumb, usually. In the grey zone! Also it depends on the source: If the source is only "on/off", then it is not a robot, merely a cofee machine.

The inputs needs to be from at least 2 different sources, have at least 6 different variations. And choise of the robot needs to be made from at least 12 different variations. And most important; The choise that the robot makes must be of a kind that can affect the next inputs to the robot.

Why 2, 6, why 12, why the need to have impact on next input? Because this is how it is! It is the universal truth!

If you have ever made a robot, you can allways hold it against the above proclaimed law of Fritsl, and you will agree; Either you did have that feeling of it being autonopmous and alive (and it passed the above test) or it did not feel autonomous, and did not pass the test, you where never in contact with creation of autonomous independance, it was not alive, it was stupid, just a complicated coffeemachine, dead.

"Is it impossible, by definition, to have a robot without sensory devices?" Yes!

"Does a pressuresensor, hooked up to the fender on a remote controlled car, that triggers a crashing noise when the car hits an obstacle, make it a robot (seeing as it now has two sources of input or senses)?" No - it has no choise, no logic other than On/off, and its output does not have the ability to change the next input, and so it is just a switch!

"b: an efficient insensitive person who functions automatically" - Hey, then I am a robot in the morning!

"2: a device that automatically performs complicated often repetitive tasks" - Hey, then I am a robot in the rest of the day!

/ Frits

I dissagree on...

Frits' definition could eliminate Industrial Robots. Those on assembly lines that follow precise movement patterns with little or no sensors and are preprogrammed by a human the first time. Some only have a pressure sensor or on/off switch to stop the Robot claw from crushing the object it just picked up. Or NASA remote controlled Lander and Rover Robots.

I aggree on...

There should be a destinction between a "Robot" [the broad meaning] and "Autonomous Robotic Life". I do agree that just RC'ing something where a Transmitter sends signals to a Reciever which is connected directly to a servo "is not a Robot". 

My 2 cents...

I would say that a "Robot" needs some form of "on-board intellegence" to interpet data. I would consider it a "Robot" if it interpets human input from RC and converts it to some action[s]. But it would be full "Robotic Life" if it read sensors and then decide whether or not to execute based on sensor data etc...