PID Controlled Bouncing Ball On Plate

[@rodrigo_dm]

gotta tell you this is the kind of thing that makes me love the hobby community. mics to get position and the pid to keep eveything working. B E A U T I F U L

[@multimoron11]

You could try using a machine learning algorithm like cross-entropy optimization and see how it compares to the PID. It's quite simple and is very applicable in PID type control systems, like balancing a stick from its bottom. Would be a cool investigation into ML.

[@carotoby]

is the sound free to use in a Techno Track ?

[@ElectronDust]

umm... Yes!

[@christopherleveck6835]

@@ElectronDust Hes shooting a pornographic movie about ping pong balls having sex....... Or is it a documentary about the mating habits of the pingus ballus?

[@victorsantos1725]

Does it use machine learning?

[@ammadejaz93]

for which purpose you want to use this sound

[@dorianmccarthy7602]

This is coming along very nicely. The thing which the system doesn't seem to take into account (Which causes the error between direction and intended direction) is the rotational direction of the ball between consecutive bounces. If that can be determined and taken into account, i think the rigs accuracy with centering/controlling the ball would majorly improve.

[@ElectronDust]

Yeah, good point. Hard to measure that though. I am thinking about going back to the camera approach for the next iteration.

[@dorianmccarthy7602]

@@ElectronDust that could be a good way of detecting rotation direction and magnitude actually.

[@davemwangi05]

@@ElectronDust is it just for fun or there's some useful purpose you're targeting ?

[@GregKrynen]

Be interesting to see if the solid arms could be replaced with filament from above, replace the square board with a ping pong paddle, and build the mics into the body and handle of the paddle with a wireless sender to the controller. A ghost paddle display.

[@JeremyCook]

Nicely done! Wow.

[@estudiantecomputerscience]

excelent work!

[@energysaver]

1 dislike from the ball

[@quaztron]

Keep it running for a few years. Does the ball wear out?

[@djlesinski7978]

I've played ping pong before. Yes.

[@MuneraPhoto]

just bring to make a small part of this project.... I play airsoft and my goal is to localise the bb's impact with microphones.... Until now, my results are unusable.... Even with your GitHub.... I suppose it is due to my microphone.... Did you have good results with any microphone or did you have to use specific ones?

[@t1mmy13]

Man this sounds like techno

[@christopherleveck6835]

Sounds like ping pong balls having sex.

[@TrueBlueKangaroo]

Brilliant! So satisfying!

[@joelgluma6125]

i am interesting about the for 4 microphone you used and hows the arduino calculate the sound of bouncing pingpong ball into the plate for impact. which software you used for visual application to draw the dots?

[@ElectronDust]

I created the visualization with Unity. Here's some more infos about how the mics work: electrondust.com/2018/07/22/stepper-juggler/

[@acorlan145]

With the issue of it being sort of off couldn't that he caused by the noises of the motors

[@TRN_SeaHag]

Weapon to surpass Metal Gear!

[@DamianReloaded]

Well that's it! Robots have officially done *everything* better than us now! **sits on the floor to watch the sunset** ^_^ Can I like this video twice? ^_^

[@no1mafiaman]

They can probably watch sunsets better than you too.

[@monkbuddy62]

shots fired

[@christopherleveck6835]

@@no1mafiaman and make them.

[@fiftystatebanana]

Do you have anymore details or images of the project? This is cool.

[@ElectronDust]

Hey Joseph, thx! Here's the code both of the Arduinos are running (the folders are "CalculatedMoves" and "MicDiffs") github.com/T-Kuhn/Stepper-Juggler I try to write up some sort of blog posty thing to explain a bit more about the thing (with pictures) within a week or so.

[@tomaszpodkowa4887]

waiting for the blog with pics

[@ElectronDust]

Here's some pics: electrondust.com/2018/07/22/stepper-juggler/#more-240

[@UkiMalefu]

the noise of the machine sounds like a voice

[@quaztron]

ya ya ya...

[@NOTAFULLUN]

You listened to it for too long and got hypnotised. I work in a factory and I asked some of the people who work on the machinery how it affects them, and one of them said he hears his machine as he is going to sleep.

[@maxk4324]

Not sure if your still working on it, but what about using piezos as contact mics underneath the plate?

[@ElectronDust]

I'm still working on it. That's a very interesting idea. I need to get my hands on some piezo sensors.

[@ariangh4358]

also use a gyro to keep the surface level when the ball isnt bouncing around

[@marshidden4352]

Arian Gh It doesn’t need one. He can just code it to reset the plate to its default position.

[@headrobotics]

Very cool 😎 do you have any plans to increase the speed and see where the system limits are? And how much it can be tweaked for response speed?

[@FatNonsense]

what do you mean with increasing the speed? The 2 things that "makes" the speed of the ball is gravity and (air) drag. If we stick with the ping pong ball and earth as testing location, there is not much we can change in aspect of "speed". Using a small steel ball may be a little bit more challenging. Or making the plate smaller, I think this is the easiest and most logical increase in challenge.

[@headrobotics]

FatNonsense wasn’t fully thought through on my part I guess; a higher table speed would result in other effects

[@FatNonsense]

I think you only want the speed of the table increased, you increase the speed capability of the servos. The result should be increasing the ability to bounce the ball higher. Higher bounce would allow derivations from the landing spot of the ball to be higher. There are other optimization one could make: - increasing the accuracy of the servos (higher angle resolution of the table tilt) thus allow for finer ball control - increasing the accuracy of the mics to allow finer tracking of the contact position of the ball on the plate

[@headrobotics]

FatNonsense what sort of triangulation algorithm is used in the software?

[@FatNonsense]

Head Robotics I don't think triangulation is used here. By using 4 mics you have 2 pairs. By comparing the trigger signal of 2 opposing mics, you will know if the bouncing spot of the ball is closer to the one or the other mic. Also if the system is precise enough you could also calculate the distance from the center with help of thr speed of sound. With each pair of mics you have 1 axis of the distance to the center spot. This data is used to tilt the plate.

[@robertcarlos5342]

Well maybe if you could make a video on HOW TO MAKE IT

[@anujgupta3326]

How much did it take to work out the correct pid variables..?

[@ElectronDust]

maybe an hour. I don't remember. I just played around with it.

[@edriskeyam726]

Nice sick bro

[@LordDecapo]

How hard would it be to make this move the ball in a set pattern

[@ElectronDust]

With the current build: nearly impossible. But that sure is something I want to do. Hope this will become possible with the next iteration.

[@LordDecapo]

@@ElectronDust sweet. I wish you luck on your next iteration!

[@ElectronDust]

Thank you @LordDecapo :)

[@uclinh697]

Where are you from?

[@victorsantos1725]

Does it use machine learning?

[@ElectronDust]

no

[@forloop7713]

Now do it with one microphone placed under the centre of the wooden square

[@conorstewart2214]

It would be difficult to get the position then you could one get the distance from the centre

[@dindafatihana3518]

Hi can i get the source code, mechanic design, and all about this project? I need for my final project in electronics engineering thank you:)

[@DowzerWTP72]

You don't deserve the degree if you're just gonna rip off someone else's work.

[@christopherleveck6835]

@@DowzerWTP72 I already called his mother.