[Dprglist] Lego, Raspberry and Python Project - Reaction Wheel Inverted Pendulum

[David P. Anderson]

Cool experiment.

Accelerometers don't help that much because when you are falling, they 
are in free fall -- as it were -- and therefore don't have any useful 
information to impart.  However, gyros tell when the body is rotating.  
That and the speed of rotation are all that's needed.

If the inverted pendulum has a moveable bottom end, like a wheel for 
example, rather than fixed in place like the example, then the gyro and 
wheel feedback together totally constrain the movement of the plant.   
Same with the classic printer carriage inverted pendulum, like for 
example this one: https://www.youtube.com/watch?v=NdbODkTNvp4

The lego experiment does seem to have to deal with lots of edge cases, 
which always raises some red flags for me.  Simpler is usually more 
robust, as a general rule.

And speaking of PID, found this great article, very complete, with some 
nifty history:

https://www.wikiwand.com/en/PID_controller 
<https://www.wikiwand.com/en/PID_controller>

thanks Karim!

dpa


On 8/21/22 6:33 PM, Karim Virani via DPRGlist wrote:
>
> */[EXTERNAL SENDER]/*
>
> https://youtu.be/WObG2LoSEwQ 
> <https://secure-web.cisco.com/16B_Ej4RXqIrh14ouP6NBo6mR_R4YDFOjBp9ZRQrDroxM9mYDkppWw8HBFDb8IRHvKuMa4GHn3IinBk2kHgEIGLz_7vBowU3xS11L2MQ-mXk0KPmX--LlwRq8dZTdicAw8oaqBoX8f4FEaxCNITE5k8HzNRPefFM-mS8jIQ7jSJt9Q5eLOnRIzyv_Yhfwm5IzExp3pW33v1cOAEjkF_krlxS6ZoTx0o0s1Jha1jHr9XXjZJvfBu_tkdppwm_ahsAenUpKi7hDBTlcXfj3S5-cDKIYpPLgsyER3SJ1favI-Lo/https%3A%2F%2Fyoutu.be%2FWObG2LoSEwQ>
>
> Just wanted to share this very nice example of deliberate and 
> experimental improvement of real world control of an inverted pendulum 
> (like walking) in this case using a reaction wheel.
>
> The number of configurations and variations tested is very cool.
>
> It also helps to clarify some concepts around dynamic balance. Like 
> accelerometers don't really help because you are not actually chasing 
> where up is, you are only chasing the orientation where the gyro tends 
> to flip direction and change the slowest, and gyro drift doesn't 
> really matter. This is well demonstrated at 11:25.
>
> I'm tempted to say some things about their usage of Angle Fix Rate and 
> whether they really got the optimal PID values, but I've had my fill 
> of PID discussions for a while.
>
> I just appreciate the thoroughness of the approach they took and how 
> well it was documented and communicated in the video edit.
>
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at lists.dprg.org
> http://lists.dprg.org/listinfo.cgi/dprglist-dprg.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.dprg.org/pipermail/dprglist-dprg.org/attachments/20220821/87024b71/attachment.htm>