@@gorak9000 I don't think so, it immediately froze. It was programmed to stop after read an unexpected pressure on its sensors. It's a common safety measure.
I understand the reasoning behind the auto stop and all... But my brain just can't help anthropomorphize it and put in that thought of "well, bugger." when I see its reaction to the sudden explosion of dice.
And this is why these things operate inside of cages... They are merciless. Edit, these are sometimes modified and used for camera motion control systems for filming special effects. occasionally they will be in close proximity to people, and while its not extremely dangerous its still quite risky as they can easily ram the camera right throigh the actor's face if the actor is off their mark or if the bot has a bad moment.
You are one of the few people who realize how dangerous robotic arms can be. They need a cage to keep people out of range of the arm swing. Idiot news people show robot arms flipping burgers in a fast food place with no guards around to keep people away. You would have to keep hiring new replacement employees to make it to the end of the shift.
@@rogermoore8977 That was true a decade ago but not today. If a robot is designed to work close to people it's classed as a co-bot. The arms can be overridden just by pushing on them, and deliberately don't have the specs of an industrial robot that can lift a car. Please don't spread paranoia, the general public are paranoid enough already. 👍
@@owensparks5013 On top of what you said (thanks for pointing it out), many have cameras placed around the robot / in the room that will detect motion (people, other robots, etc) and will make the robot slow down / shut off if its in proximity of other people / robots. I feel safe around a UR5 flipping burgers, its slow and weak, it can do me less harm than a person flipping burgers if it hit me comapred to a person hitting me with an albow.
@@harounben342 nope, for this you need colaborative robots. Most robot companies also produce cobots. They are made under different regulations and standards.
@@harounben342 Unfortunately all computer numerical control machines must be STILL enclosed for safety of humans. Also, fortunately for us, They are so powerful and they don’t know that YET!!!???
Sitting here watching this it looked almost like with the last dice it’s stacked on. It made the stack really wobble and I can’t help but thinking if the arm just got pissed off at it moving so much and decided to take it out
This was my thought exactly. The robot was so close, saw it wobble, got pissed, and said NO! If this thing falls it's going to be me who does it. Sort of like the robot's version of flipping the table.
Reminds me of the time during my uni course when I was programming something similar to draw pictures (albeit a smaller desktop one) and accidentally put an extra 0 in the speed setting... (it wasn't bolted down properly and moved so fast it managed to flip itself off the desk!)
Haha I have a similar project right now and I accidentally made a for loop that started with the robot at the resting position (on the table). So every time the loop finished, the robot would smack itself onto the metal table. Luckily it didn’t break.
Use the super secret keyboard keys to step through the video one frame at a time. "Period" is forward one frame, "comma" is back one step. This way you can see that the algorithm was flirting with success from the beginning, as the bot head moved downwards, it was always coming very close to the inside of the vertical stack. When the last two were stacked, they had a bit of an "inward" lean to them, and the arm was just "kissing" the edges. That last one finally came a little too close to the inside edge of the die, and POW!
There’s also a bit of sway that was happening, whether from the super tall dice stack just being unstable or from the movements of the arm making the table move a little due to the abruptness and speed of the arm’s motion’s inertia transferring into the floor and to the table. Probably a combination of the two caused the movement of the dice stack into the arm path. Looks like maybe it’s on the second floor? Probably not as rigid as being mounted on ground level concrete would be.
@@zaprodk They can and do sense crashes. It does so by sensing an increase in current in any of the six motors. He is a pretty smart guy and knows approx. how much current should be drawn when doing specific movements at specific accelerations/speeds. How sensetive it should be is also adjustable (and i guess that you also can disable it, not sure).
I think there’s something human about mechanical arms that makes them feel familiar and safe, but it’s really important that everyone knows this is not the case, they are absurdly dangerous and, when being used experimentally, very temperamental. At an apprenticeship thing I visited they had one and on the floor there was a red tape boundary around it at its full extension, which had been aptly dubbed the “Kill-Zone”…
I think that was a hardware/software safety feature, if too much pressure is applied in any specific direction, it will activate the E-stop, immediately halting the robot from whatever it was doing. The robot is just a bunch of metal and plastic held together by very powerful motors and checked by sensors, things can go very wrong very quick without redundancy and safety compliance
@@jpsalis Not really a pressure difference as it doesn't seems to have any pressure sensor but rather a position deviation (between the current one and the calculated one from the interpolator)
@@Difool80 Something like that, since each motor has feedback it is posible so measure the amount of torq aplied for each of them, which can be translated to the force aplied for the robot on its tool if it goes too high it returns an error and the robot stops
Why does it do this quick left and right detour instead of placing it directly? Does it measure something? And does it work based on absolute pre programmed values like length of the cubes and simply does cubeheight×number or does it work more dynamically?
woah yeah, I would not have caught that very physical runtime error, someone's gotta adjust the design there lol also that bot was near full extension, no way it could've done much more than that without the stack being closer or the table being closer
@@Quantum-Bullet indeed, it measures the current in the motors and compares to the computated theoretical currents. if they are too different, it will automatically shut off.
"it's the additions that scares me" - guy in the background "it's gonna kick it" - guy holding the camera "Yeah that's right hahahah" -guy in the background They were worried about it happening. The robot has to change it's program for each dice. it has to add the length of a die each time. That's probably why the guy holding the camera didn't boldge. He knew he was safe. P.S. J'adore ce que vous faites! (future ingénieur roboticien)
I've baby-sat a welding robot for years. I loaded parts for it and also jigged up my own parts to weld while it worked. My average was 1.5 finished parts for every 1 part the robot made.
Robot is only cheaper if it doesn't need a babysitter, otherwise you pay for an employee AND a robot. The math says it saves money in the long run and ups productivity by 66% over just an employee though.
@@CLove511 one employee for like 10 or more robots. Math says you can buy another robot and staff it by the same person, while minimally increasing overhead.
But the robot can continually do 1 part ever run where as you'll not keep up, over a certain given time.... Other side of the coin, the robot is useless without you.
Feels like the robot I maintain. What a horriffic beast leading its own life once in a wile. Also, the use of newer LMS Lidars made our little wrecker function a lot better. His behaviour course has deffenitely worked. Altough, its still a mad cow in what it does somedays xD It has no specific motions it repeats and picks different kind of farmats of objects in a pickup belt. It sorts them and try placing them as efficient as possible. But operators need to keep an eye on it, since a child or 3 could do better, in ordering the sorting proces, and deffenitely will not break a crate either, or even smash its pickoff belt. Which is deffenitely programmed as a dead zone xD
Irb 2400? I have one doing diamond deburr.its basically a half scale Irb4400 and one of the best built Abb robots out there due to the link arm construction the are very rigid and virtually indestructible.I had a tech slam it into a IRB 6640 and it didn't break anything on the robot.We have had a IRB 4400 running since the 90s and it was in the absolute worse foundry conditions and lasted until I removed it this year.
The collision is a programming error(human) re: changing enviromental constraints with the target, but more significantly is the HORRENDOUS axis "zeroing" for each cycle AND axis!!!!! A human would use two hands, in a cyclic and complex manner and in a fraction of the time - humans rock!!!
By the looks of it, they hit the safety glass in about 1/6 of a second. The glass seems to be about 1 meter away from the dice, so they flew approximately 6 m/s.
And if it hit the glass in 1/12th of a second then it would be 12m/s and if it hit the glass in 1/3th of a second then it would be 3m/s…. Kinda inaccurate to guess a speed this way “by the looks of it”.
I am pretty sure what they do here is they minimize backlash. The joints are geared and therefore have a little bit of backlash which adds up a lot in a long chain like a robot arm. The worst thing is, depending on which direction each joint was previously rotating in, the tooth of the gears are engaged in different directions. Because of that, the robot will end up at different actual positions depending on whether it arrived at that same target position from one side or the other. Therefore it is best to move the joints in a particular direction to align the gears before performing the placement from another direction. There will still be backlash, but at least it is always the same, so the repeatability is better (i.e. the robot might not arrive at the exact requested position, but at least it will be off by the same amount in the same direction every time). This is common to all robot arms. Some do have very small backlash but are also way more expensive. It all depends on the use case. As you can see here, if done right, you can get pretty high repeatability even with a (presumably) not backlash minimized robot.
The replies to another comment suggest that it since this was a demo, it was likely just to show off machine’s fast movement and precision even afterwards, like a magician waving their props around during an act. That seems more likely to me, because there would be no need to go so far or so quickly just to take up backlash (which I expect would be minimal anyway), though I am a non-expert.
@@PaulFisher Well, there is no way to be sure except with a statement from the engineers that made the program. I can only tell you (as an expert), that this type of movement is common to improve repeatability. Backlash in robot arms is also a very big deal and you can't expect it to be minimal by default.
Quite often they move to a known "home" position before starting each individual procedure. So from that home position it moves "distance" + "X" to each new placing position. I worked on these for years and never had real issues with backlash. Once they got that bad they would be taken out of service and rebuilt with new parts.
I can imagine some movie scene in which a robotic crime lord stacks dice on its desk, only to snap them forward like this to dispose of an incompetent human thug.
Where I work we had a collision with a Kuka robot. Lifted an 80,000 pound 110ft long tool into the air and took a 5” chunk out of an inch and a half thick carbon fiber airplane part. Robot was fine. Several million dollars in damage because the programmer missed a “-“
VERY good catch! I'm assuming that was the reason for the crash; had it been aligned it may have gone flawlessly. Now the engineers/programmers have to keep more safety distance due to uncertainty positioning from the wobbles.
This was not a collision. The two of dice became unstable and fell in the direction of the robot. Always nice seeing the repeatability of these machines. 😍👍👍👍👍👍
And this why you don't go near robit arms in service. They will hurt or even end your life painfully. They may be elegant but the torque is enough to break many things
"Eventually," you may not realize the reason why that robot has always been surrounded by such a huge cage. Machines have been killing people for centuries.
Notice the left and right movement stayed at the same height . Should have been programmed to go to the new height before doing the left and right move
The arm was distracted doing the fancy left right dance before placing the cube, it forgot to place the cube, pull back, then go down to pick up the next cube. It hits it head.
They didnt take into account the height of the dice tower when programming the waypoints. It is a linear interpolation from the top of the tower to the next die. The angle of the edge of the imaginary triangle gets shallower and shallower each die.
@@nrdesign1991 Yeah that's kinda my thought. It's programmed to retreat a certain distance, but as the arm goes above level it needs to retreat more as the arc of rotation brings it back close and in this case contacts the stack.
I’m taking an inverse kinematics class at university right now and we all have to do a final project using inverse kinematics on a $30,000 robot arm. Figuring out how to control it was mostly trial and error so everyone’s code has inadvertently smacked the arm onto the metal table at full speed at least once. It’s taken it like a champ though. Luckily everyone knows to stand back, so no one’s been injured.
My mom bought me a "Super Armitron" when I was a kid and it was a cool concept, but with a very weak mechanism. From time to time I think about how much I want a robotic arm.. would I build it? Would I be able to find a use for it? What if I stuck it on wheels? I think a robotic arm on wheels would be cool. Aerial drones are heavily regulated, but what about ground drones? That would be basically be like an RC car. Then again, they may just adapt the law which is used with drones in which the drone must be within "line of sight" of the operator.
The tormach arm is supposed to launch at $20k iirc That plus maybe a car and some electronics and you're off to the races, becoming the next supervillain in no time.
Just machines folks, these things don't "make mistakes" it's always the operator. One day when humans finally achieve general artificial intelligence, we may be able to utilize such AI to complete tasks with these machines. We won't have to program anything, just let the AI observe the process and end result and it'll do it quicker and more efficiently than this programming we see here (doing a little dance before placing the block). But till then, it's silly mistake ridden humans. Humans are simply a stepping stone to a new form of exotic life.
It actually wasn't the robot making a mistake. It was because the tower was tipping over slightly, so the tower wasn't aligning with where the robot thought the tower would be.
it's actually the programmer's fault for not giving more clearance. I had expected an eventual collision when I saw it making the exact same linear moves as the tower grew higher.
@@jungy537 It looks like a demo booth, so it's possible all the extra movements and the tiny clearance to the arms path was all down to the programmer being told to "Make it attention grabbing".......... Which in a way, they did. :)
@@Reman1975 I was one of the 3 in that room and you pretty much nailed it. We were first year students fooling around to make a cool video. Did not expect THAT to happen though.
CNC machines always look elegant, graceful and precise, right up until the moment that something like this happens.
i wouldn't exactly call the noise graceful lol
They are only as elegant as they are programmed to be. Rubbish in, rubbish out.
Kind of like the 3D printers too, they also wait till the end to have a catastrophic failure
"WHO FUCKED WITH THE OFFSET?"
bad programming. human error
That actually scared the fuck out of me😂
Me too lmao
@@Felix-vb3eu same dude
fr i almost threw my phone.
It caught me so off guard lmao
Yep me too damn
Kudos to the cameraman for not making his cellphone/camera fly off as well. xD
the cameraman is another robot
His underwear, not so lucky
probably because this was a repeat of what already happened :)
@@dsfs17987 yep, …almost like they were expecting it!
lol
I like how it freezes at the end like "... Shit."
Or more like someone hit the e-stop button
@@gorak9000 Thats too bad, I really wanted to see it continuing by just dropping the dice from the height
@@gorak9000 I don't think so, it immediately froze. It was programmed to stop after read an unexpected pressure on its sensors. It's a common safety measure.
I understand the reasoning behind the auto stop and all... But my brain just can't help anthropomorphize it and put in that thought of "well, bugger." when I see its reaction to the sudden explosion of dice.
@Tom R Why are you grabbing crazy long unusual words from the internet to boast your IQ here?
And this is why these things operate inside of cages... They are merciless. Edit, these are sometimes modified and used for camera motion control systems for filming special effects. occasionally they will be in close proximity to people, and while its not extremely dangerous its still quite risky as they can easily ram the camera right throigh the actor's face if the actor is off their mark or if the bot has a bad moment.
Yes , it is always fun to see the robot having a ...ah fcuk it moment as throws away the 50kg bags as nothing......
yeah
they will move through you with ease
You are one of the few people who realize how dangerous robotic arms can be. They need a cage to keep people out of range of the arm swing. Idiot news people show robot arms flipping burgers in a fast food place with no guards around to keep people away. You would have to keep hiring new replacement employees to make it to the end of the shift.
@@rogermoore8977 That was true a decade ago but not today. If a robot is designed to work close to people it's classed as a co-bot. The arms can be overridden just by pushing on them, and deliberately don't have the specs of an industrial robot that can lift a car. Please don't spread paranoia, the general public are paranoid enough already. 👍
@@owensparks5013 On top of what you said (thanks for pointing it out), many have cameras placed around the robot / in the room that will detect motion (people, other robots, etc) and will make the robot slow down / shut off if its in proximity of other people / robots. I feel safe around a UR5 flipping burgers, its slow and weak, it can do me less harm than a person flipping burgers if it hit me comapred to a person hitting me with an albow.
I was wondering why they needed those screens...
Safety
The safety barriers are the must in a vicinity of working robots.
so you can't work side by side with these powerful monsters yet?
@@harounben342 nope, for this you need colaborative robots. Most robot companies also produce cobots. They are made under different regulations and standards.
@@harounben342 Unfortunately all computer numerical control machines must be STILL enclosed for safety of humans. Also, fortunately for us, They are so powerful and they don’t know that YET!!!???
Never knew a stack of dice can explode
How about a stack of paper? ruclips.net/video/qTz7aKEJLV4/видео.html
Hydraulic forces can make a lot of things explode.
Sitting here watching this it looked almost like with the last dice it’s stacked on. It made the stack really wobble and I can’t help but thinking if the arm just got pissed off at it moving so much and decided to take it out
This was my thought exactly. The robot was so close, saw it wobble, got pissed, and said NO! If this thing falls it's going to be me who does it. Sort of like the robot's version of flipping the table.
When your release meets real world conditions.
Reminds me of the time during my uni course when I was programming something similar to draw pictures (albeit a smaller desktop one) and accidentally put an extra 0 in the speed setting... (it wasn't bolted down properly and moved so fast it managed to flip itself off the desk!)
Haha I have a similar project right now and I accidentally made a for loop that started with the robot at the resting position (on the table). So every time the loop finished, the robot would smack itself onto the metal table. Luckily it didn’t break.
Use the super secret keyboard keys to step through the video one frame at a time. "Period" is forward one frame, "comma" is back one step. This way you can see that the algorithm was flirting with success from the beginning, as the bot head moved downwards, it was always coming very close to the inside of the vertical stack. When the last two were stacked, they had a bit of an "inward" lean to them, and the arm was just "kissing" the edges. That last one finally came a little too close to the inside edge of the die, and POW!
1. Amazing, did not know these keys did that
2. I read that in Sam's voice, kudos!
What algorithm?
There’s also a bit of sway that was happening, whether from the super tall dice stack just being unstable or from the movements of the arm making the table move a little due to the abruptness and speed of the arm’s motion’s inertia transferring into the floor and to the table. Probably a combination of the two caused the movement of the dice stack into the arm path. Looks like maybe it’s on the second floor? Probably not as rigid as being mounted on ground level concrete would be.
I saw that my first go, and I didn’t need to slow it down :-P
@@jensaugust743 the one that controls the robot
I like how it stops moving out of frustration. Just like a real person who thinks about what he has done
It stops because the operator stopped it. It has no feedback mechanism, so it cannot sense that something went wrong.
@@zaprodk They can and do sense crashes. It does so by sensing an increase in current in any of the six motors. He is a pretty smart guy and knows approx. how much current should be drawn when doing specific movements at specific accelerations/speeds. How sensetive it should be is also adjustable (and i guess that you also can disable it, not sure).
@Tom R If you are standing with the controller in hand with your finger on the e-stop, yes you can.
@@zaprodk I think it was a joke…
I think i poo my self a little. I was expecting it to just knock them over not blow them up lol.
I think there’s something human about mechanical arms that makes them feel familiar and safe, but it’s really important that everyone knows this is not the case, they are absurdly dangerous and, when being used experimentally, very temperamental. At an apprenticeship thing I visited they had one and on the floor there was a red tape boundary around it at its full extension, which had been aptly dubbed the “Kill-Zone”…
Wtf really? Can't believe that. We have Cages with a lot of sensors around them.
Loved how it’s CPU instantly blew up as soon as it exploded the dice. Robbie the robot has a new friend. “Warning,warning!!” 😂
I think that was a hardware/software safety feature, if too much pressure is applied in any specific direction, it will activate the E-stop, immediately halting the robot from whatever it was doing. The robot is just a bunch of metal and plastic held together by very powerful motors and checked by sensors, things can go very wrong very quick without redundancy and safety compliance
@@jpsalis Not really a pressure difference as it doesn't seems to have any pressure sensor but rather a position deviation (between the current one and the calculated one from the interpolator)
@@Difool80 Something like that, since each motor has feedback it is posible so measure the amount of torq aplied for each of them, which can be translated to the force aplied for the robot on its tool if it goes too high it returns an error and the robot stops
Why does it do this quick left and right detour instead of placing it directly? Does it measure something? And does it work based on absolute pre programmed values like length of the cubes and simply does cubeheight×number or does it work more dynamically?
It does it for visual show.
Guessing it's just written code using inverse kinematics. Nothing fancy other than math. The left and right detour is just for show
Programmer used starting position for every cykle. No reason to do it but in real use in industry we don't do it that way.
Possibly calibrating the position every time
I think they do that to prove accuracy. After all those movements it still knows exactly where to place the dice. A little show-off in accuracy.
it's been 5 years, I hear it can stack up to 19 dice now before catastrophic failure.
woah yeah, I would not have caught that very physical runtime error, someone's gotta adjust the design there lol
also that bot was near full extension, no way it could've done much more than that without the stack being closer or the table being closer
I like how it came to a sudden stop like "Phauucckkkkkk...." 😲
Hahaha was thinking the same thing 🤣
Probably automatic safety procedure due to unexpected forces.
@@Quantum-Bullet indeed, it measures the current in the motors and compares to the computated theoretical currents. if they are too different, it will automatically shut off.
"it's the additions that scares me" - guy in the background
"it's gonna kick it" - guy holding the camera
"Yeah that's right hahahah" -guy in the background
They were worried about it happening. The robot has to change it's program for each dice. it has to add the length of a die each time. That's probably why the guy holding the camera didn't boldge. He knew he was safe.
P.S. J'adore ce que vous faites! (future ingénieur roboticien)
I've baby-sat a welding robot for years. I loaded parts for it and also jigged up my own parts to weld while it worked. My average was 1.5 finished parts for every 1 part the robot made.
If only everyone had Your skillz
robotwasprobablycheapertho
Robot is only cheaper if it doesn't need a babysitter, otherwise you pay for an employee AND a robot. The math says it saves money in the long run and ups productivity by 66% over just an employee though.
@@CLove511 one employee for like 10 or more robots. Math says you can buy another robot and staff it by the same person, while minimally increasing overhead.
But the robot can continually do 1 part ever run where as you'll not keep up, over a certain given time.... Other side of the coin, the robot is useless without you.
looks so delicate but then it does something like this and you realize it could rip you limb from limb
Feels like the robot I maintain. What a horriffic beast leading its own life once in a wile. Also, the use of newer LMS Lidars made our little wrecker function a lot better. His behaviour course has deffenitely worked. Altough, its still a mad cow in what it does somedays xD
It has no specific motions it repeats and picks different kind of farmats of objects in a pickup belt. It sorts them and try placing them as efficient as possible. But operators need to keep an eye on it, since a child or 3 could do better, in ordering the sorting proces, and deffenitely will not break a crate either, or even smash its pickoff belt. Which is deffenitely programmed as a dead zone xD
Irb 2400? I have one doing diamond deburr.its basically a half scale Irb4400 and one of the best built Abb robots out there due to the link arm construction the are very rigid and virtually indestructible.I had a tech slam it into a IRB 6640 and it didn't break anything on the robot.We have had a IRB 4400 running since the 90s and it was in the absolute worse foundry conditions and lasted until I removed it this year.
yep, it's quite amazing huh
The collision is a programming error(human) re: changing enviromental constraints with the target, but more significantly is the HORRENDOUS axis "zeroing" for each cycle AND axis!!!!!
A human would use two hands, in a cyclic and complex manner and in a fraction of the time - humans rock!!!
You can tell the robot is more precise than the dice it stacks...
Damn these crashes are flooding my recommended… not complaining!
By the looks of it, they hit the safety glass in about 1/6 of a second. The glass seems to be about 1 meter away from the dice, so they flew approximately 6 m/s.
🧢
And if it hit the glass in 1/12th of a second then it would be 12m/s and if it hit the glass in 1/3th of a second then it would be 3m/s…. Kinda inaccurate to guess a speed this way “by the looks of it”.
@@Engineer9736 exactly, bet he guessed the ‘1/6 of a second’ with his ears
Thought I’d fallen for some click bait or “collision” had a different meaning in this context, then all of a sudden DIE!
Robot: Right I’ve had enough of this!!
**Proceeds to head it the shit out of the pile** 😂
Why it makes the same left-right movement every time it picks a dice up?
I am pretty sure what they do here is they minimize backlash. The joints are geared and therefore have a little bit of backlash which adds up a lot in a long chain like a robot arm. The worst thing is, depending on which direction each joint was previously rotating in, the tooth of the gears are engaged in different directions. Because of that, the robot will end up at different actual positions depending on whether it arrived at that same target position from one side or the other. Therefore it is best to move the joints in a particular direction to align the gears before performing the placement from another direction. There will still be backlash, but at least it is always the same, so the repeatability is better (i.e. the robot might not arrive at the exact requested position, but at least it will be off by the same amount in the same direction every time). This is common to all robot arms. Some do have very small backlash but are also way more expensive. It all depends on the use case. As you can see here, if done right, you can get pretty high repeatability even with a (presumably) not backlash minimized robot.
The replies to another comment suggest that it since this was a demo, it was likely just to show off machine’s fast movement and precision even afterwards, like a magician waving their props around during an act. That seems more likely to me, because there would be no need to go so far or so quickly just to take up backlash (which I expect would be minimal anyway), though I am a non-expert.
@@PaulFisher Well, there is no way to be sure except with a statement from the engineers that made the program. I can only tell you (as an expert), that this type of movement is common to improve repeatability. Backlash in robot arms is also a very big deal and you can't expect it to be minimal by default.
@@mariodekena1005 interesting, thank you!
Quite often they move to a known "home" position before starting each individual procedure. So from that home position it moves "distance" + "X" to each new placing position. I worked on these for years and never had real issues with backlash. Once they got that bad they would be taken out of service and rebuilt with new parts.
This robot's priceless reaction like "oh poop"
programmer's fault
scared the shit out of me, that was way more violent than i expected
1:10 my reaction when I know I fu*ked things up 😂 poor guy who will collect all this mess it did.
1:10 that face was priceless
I can imagine some movie scene in which a robotic crime lord stacks dice on its desk, only to snap them forward like this to dispose of an incompetent human thug.
I wonder why the RUclips algorithm but this cool little vid into my feed.
The robot: we were on the verge of greatness, we were this close
And that's what the fence is for lol
why is the arm doing this left right jiggle every time? is this calibrating?
Where I work we had a collision with a Kuka robot. Lifted an 80,000 pound 110ft long tool into the air and took a 5” chunk out of an inch and a half thick carbon fiber airplane part. Robot was fine. Several million dollars in damage because the programmer missed a “-“
Notice the last cube to be successfully placed is out of alignment
VERY good catch! I'm assuming that was the reason for the crash; had it been aligned it may have gone flawlessly. Now the engineers/programmers have to keep more safety distance due to uncertainty positioning from the wobbles.
It tried to place the last few dice too far down.
This was not a collision. The two of dice became unstable and fell in the direction of the robot. Always nice seeing the repeatability of these machines. 😍👍👍👍👍👍
And this why you don't go near robit arms in service. They will hurt or even end your life painfully. They may be elegant but the torque is enough to break many things
Eventually, there will be some robot - humans fatalities for sure. We have to be always very careful around these machines!!!
"Eventually," you may not realize the reason why that robot has always been surrounded by such a huge cage. Machines have been killing people for centuries.
@@ButterfatFarms Thanks to let me know.
Skynet: I will destroy humans
Me: Idiots wrote your software, you have 0 chance.
why does it do those horizontal movements before starting to move upwards? they doesn't look necessary to me...
To prove it's accuracy
@@Nortem Not sure how they expected it to finish. The arm was getting pretty close to the limits of its travel.
Next video: "robot surgeon"... this might be interesting
the robot surgeon would not be able to make such large movements, but nice try
Notice the left and right movement stayed at the same height .
Should have been programmed to go to the new height before doing the left and right move
why do you make it go back to the reset point before setting them down? bad code? or intentional?
The arm was distracted doing the fancy left right dance before placing the cube, it forgot to place the cube, pull back, then go down to pick up the next cube. It hits it head.
They didnt take into account the height of the dice tower when programming the waypoints. It is a linear interpolation from the top of the tower to the next die. The angle of the edge of the imaginary triangle gets shallower and shallower each die.
@@nrdesign1991 Yeah that's kinda my thought. It's programmed to retreat a certain distance, but as the arm goes above level it needs to retreat more as the arc of rotation brings it back close and in this case contacts the stack.
why people say they were scared of the robot, it looked a bit of Wall-e, working alone, poor guy, and made a mistake at the end.. :-(
Humans did the mistake not the robot
The robot did precisely what it was ordered to do
Missed it by _that_ much...
Robot: Oh no... Maybe they didn't see that! No disassemble!
Number five is alive!!
That was strangely funny
Robot was like ahh shit i fucked it up then acts innocent and stays at 1 place 😂🤣
Robot be like :
"looke my " -
POW!!!!
Its look of disappointment at the end... 😔 I feel you bro
I’m taking an inverse kinematics class at university right now and we all have to do a final project using inverse kinematics on a $30,000 robot arm. Figuring out how to control it was mostly trial and error so everyone’s code has inadvertently smacked the arm onto the metal table at full speed at least once. It’s taken it like a champ though. Luckily everyone knows to stand back, so no one’s been injured.
Top manager, probably...
I havent screamed like that since my brother hit me with the scary maze game as a kid
You scared me too! LOL
Honnêtement avec la petite dance je m'attendais à pire. Mais pour un premier programme c'est très bon
I flinched
My mom bought me a "Super Armitron" when I was a kid and it was a cool concept, but with a very weak mechanism. From time to time I think about how much I want a robotic arm.. would I build it? Would I be able to find a use for it? What if I stuck it on wheels?
I think a robotic arm on wheels would be cool. Aerial drones are heavily regulated, but what about ground drones? That would be basically be like an RC car. Then again, they may just adapt the law which is used with drones in which the drone must be within "line of sight" of the operator.
The tormach arm is supposed to launch at $20k iirc
That plus maybe a car and some electronics and you're off to the races, becoming the next supervillain in no time.
Funniest thing I've seen all day
It was getting wobbly at the end. It was like a kid that knows he's going to fail at building the tower. So, he just knocks it over.
A stark reminder that it's probably best to do the simple things by hand.
Robo woke up and chose violence
You can see the collision detection working and immediately stopping the robot after making the dice fly
LOL...did you use a MoveJ there?
I want to see Kuka coasters on the production line, with riders along for every step of the process.
The jumpscare is at 1:10 , you have been warned , dont die of a heart attack
Just machines folks, these things don't "make mistakes" it's always the operator.
One day when humans finally achieve general artificial intelligence, we may be able to utilize such AI to complete tasks with these machines.
We won't have to program anything, just let the AI observe the process and end result and it'll do it quicker and more efficiently than this programming we see here (doing a little dance before placing the block).
But till then, it's silly mistake ridden humans.
Humans are simply a stepping stone to a new form of exotic life.
Better stack than i could do
The very moment Skynet became self aware...
期待を裏切る終わり方!!!
Did you hear the robot scream afterwards "I ain't doin' this no more, you lot can get stuffed".
Wasnt expecting that ^^
Exciting video
I love how the arm froze after everything went to hell
That actually scared the crap out of me
Try to imagine an surgery robot fail :-)
That robotic arm seems to be suffering from OCD...the way it moves right to left to right before placing each piece!
Фиаско!
Dude I got an heart attack from that. Jesus!
I see that your vidoeo has fallen in the good graces of the algorithim gods
The worrying thing about AI is if it doesn’t get it right first time it may well just destroy its creation 😳
And it's back to Robot Prison for you then...
0:55 anxiety level begin to rise exponentially
Yep, programmer didn't thought about the very simple thing that the angle will get smaller the higher up it will get.
Daumn!! I gotta make one of those!!
Now run programme 2 - picking the dice up again ;)
Looks like it's performing sleight of hand between the stacks
That thing looks like it has a mind
Then It shouts: drop your gun. You have 5 seconds left...4...3...2...
Ah no, you have 0.0025 seconds left.
Or what?? You gonna hit me with these dices? Not impressed!
Robot: 1:10
Epic
It actually wasn't the robot making a mistake. It was because the tower was tipping over slightly, so the tower wasn't aligning with where the robot thought the tower would be.
it's actually the programmer's fault for not giving more clearance. I had expected an eventual collision when I saw it making the exact same linear moves as the tower grew higher.
@@jungy537 Yes, exactly. This is where experience comes in.
@@jungy537 maybe it will be possible for the program calculate clearance itself in 3d simulation
@@jungy537 It looks like a demo booth, so it's possible all the extra movements and the tiny clearance to the arms path was all down to the programmer being told to "Make it attention grabbing".......... Which in a way, they did. :)
@@Reman1975 I was one of the 3 in that room and you pretty much nailed it. We were first year students fooling around to make a cool video. Did not expect THAT to happen though.
But can it stack a standard set of D&D dice.