Are they really robots?

[JamesCH95]

I remember being in 3rd grade and talking to my parents about the definition of a robot. What we came up with has agreed with just about everyone I've talked to about it:

A robot is a machine that is controlled through a programmable computer, that is capable of autonomous or pre-programmed behavior, but it can also be "tele-operated" (to borrow a FIRST term).

I would argue that FRC does build robots, even if they are not fully-autonomous, machine-learning, bajillion-integrated-sensor robots that require tens or hundreds of thousands of dollars to build and years of development to fine-tune.

[Mark McLeod]

We teach high school students to build Cyborgs.

"self-regulating human-machine systems" - 1960 Clynes & Kline.

Our human-machine entities are joined at the joystick...

[davidthefat]

I always thought robots were reprogrammable machines... But TBH, the ones we make seem like just complex RC machines

[Ty Tremblay]

In the WPI Robotics Engineering Department, the general (read: boiled down) definition of the robot is a machine that can sense the world, make "decisions", and act upon the world.

IMHO FIRST robots are true robots in that sense during the :15 of autonomous. However, many robots continue to use sensors throughout teleoperated. So, while they're not entirely autonomous, they're still performing autonomous tasks throughout the entire match and are thus robots.

[Ether]

Originally Posted by JamesCH95

A robot is a machine that is controlled through a programmable computer,...

So if it's controlled by a computer that's not programmable, it's not a robot?

[Ether]

Originally Posted by Rich Kressly

I remember being in a similar, and equally amusing, conversation about what "music" is when Rap was becoming more prevalent in the mainstream about 18 years ago.

Reminds me of the arguments I used to hear about "What is art?"

I came up with my own personal definition, which has served me quite well over the years: (scroll down)

If I can do it, it's not art.

[Ed Law]

It may be easy to look at this from another point of view for people who thinks that FIRST Robots are just machines during teleoperated mode and not robot. Is it just an expensive remote controlled vehicle?

What is the definition of a remote control car? My definition and the common definition is that a remote control car will do exactly what you tell it to do. If you move a joystick, it will move. It will never disobey you unless the battery runs out. I would argue that a FIRST robot with sensors is not a remote control vehicle. It is because even in teleoperated mode, it may behave differently than what your joystick says. It just takes your input as a suggestion and together with other input, makes a decision what to do. It does that to protect itself from damage, to stay within the envelop/rules or for whatever reasons that it was programmed to do. You may think that you are remotely controlling the robot but you are not. Since it exhibits artificial intelligence, it should be classified as a robot.

On the other hand, a FIRST Robot that does not use sensors during teleoperated mode is a remote control vehicle.

[MrForbes]

Hmmm....I wonder if the joystick is a sensor?

[Daniel_LaFleur]

Originally Posted by Taylor

According to a display I saw at the Indiana State Fair two months ago, robots have four distinct components:

1. Sensors
2. Controller/Program
3. Kinematics/Mechanisms
4. Actuators/Motors

.

Sounds like humans are robots.
1. Sensors -- Eyes, nose, etc
2. Controller/program -- Brain/teachings
3. Kinematics/Mechanisms -- Arms,Legs,etc
4. Actuators/Motors -- Muscles

[RoboDesigners]

I think I'm going to bring that one up with my biology teacher... we're discussing the definition of life. I think as technology progresses, definitions might have to change.

[dlavery]

Back in 1976, Tom Sheridan (Emeritus, Prof Eng & Applied Psychology [Mech. Engr.], Prof Aero & Astro, Massachusetts Institute of Technology -- and former office-mate of Dr. Woodie Flowers) defined the term "Telerobot" as a device that exhibited the capabilities for either teleoperated control, autonomous control, or shared supervisory control between the two modalities (*1). Later, in 1992 (*2), he refined the definition with a clarification of supervisory control as "in the strictest sense, supervisory control means that one or more human operators are intermittently programming and continually receiving information from a computer that itself closes an autonomous control loop through artificial effectors to the controlled process or task environment."

Based on both the strict interpretation, and the intent, of Sheridan's terms, it seems that current FRC machines perfectly satisfy the definition of "telerobots." I would have no problem at all using that term to reference the devices we build. The only real implication of this is that the "FIRST Robotics Competition" ("FRC") will have to be renamed "FTC." The current "FTC" will have to find another acronym. I dunno, perhaps "FVC"?

-dave

*1 - NASA Telerobotics Program Plan, Office of Aeronautics and Space Technology, NASA Headquarters
*2 - Telerobotics, Automation, and Human Supervisory Control. MIT Press, Cambridge. p. 1. ISBN 9780262193160.



.

[MagiChau]

Originally Posted by Daniel_LaFleur

Sounds like humans are robots.
1. Sensors -- Eyes, nose, etc
2. Controller/program -- Brain/teachings
3. Kinematics/Mechanisms -- Arms,Legs,etc
4. Actuators/Motors -- Muscles

Don't some people say humans are living machines?

[RoboMaster]

I would say we build robots because they at least think for themselves to some degree. Even if you just press a button or move the stick on a joystick, (implying your judgment and thought on it) the robot has to make sense of it and act accordingly. I would say the lowest level definition of a robot would be that it takes input (which could just be its own programming), interprets it, and acts using some form of kinematics/actuators (no kinematics=computer).

It certainly wouldn't be a robot if it was completely mechanical parts. That's a machine or tool. I'd think that would be too direct to be a robot. Likewise, many people think RC cars are not robots, their level of control is too direct. I would agree if the RC car was very simple and just used circuits to transform the radio signals into power for motors. But if the RC car was complex - a glorified one - I would call that a robot.

Another aspect to remember is the "common conception" of robots. Things are labeled as robots if they have robot-y aspects like: looking like humans, small vehicles, arms, actuators, and if they do cool things. A washing machine could technically be a robot, but it doesn't have any of these aspects and its purpose is more like a machine, so that's what it's called. Our robots are very much "roboticy," as the common person would think.

Here's an idea: make a list with various levels of "robot-ness." Where does it stop and turn into something else?
-Sentient humanoid
-Sophisticated autonomous robot system (think car manufacturing)
-Simple autonomous robot system (think hobby robot)
-FIRST Robot
-Complex, semi-autonomous, semi-direct-control "RC car"
-very direct RC car
-"car" with two motors that have long wires going to two switches that are held in operator's hand. (RC car without the RC, plus it's simple)
-electric drill
-computer (just software)
-mechanism or linkage (just hardware)

Anyone think of a different kind of list or different things to add?

[Ether]

Originally Posted by RoboMaster

It certainly wouldn't be a robot if it was completely mechanical parts.

Why not ? You can make sensors, actuators, power sources, and computers from completely mechanical parts.

[RoboMaster]

But then they would be more mechanisms or machines. I know what can be done with just mechanics, but I'd be pretty impressed if you were able to make a full fledged "robot" like a hobby robot out of just mechanical parts. Even RC cars have electronics for the radios and stuff.

I'll even give a link to something (mostly) mechanical: A mechanically programmed Lego car: http://tinkernology.blogspot.com/201...-computer.html I would still call this a robot because: it has multiple robot-y aspects as described above, and I would count this unique way of programming still programming and "deciding for itself."

What I'm talking about is like if you scrapped all electronics on your FRC robot, had a really long hand crank with a universal joint that gives power to the wheels, and a similar hand crank for directly altering the steering. That's more of a very fancy mechanism.

But I guess people might want to call even that a robot because of its robot-y qualities. I guess I'd be ok with that if it really is robot-y enough. Times change and the term robot is loosely applied, like the dishwasher analogy in reverse. That's why I include robot aspects as part of my personal definition above.



Here's something I thought of: on Mythbusters they often build "robots/mechanisms" to run their experiments. They call them robots. Are they really robots? Would that apply?