Inverted Pendulum - Creative Programming!

[brianburke@bwc]

Check out this new video of a linear actuator used as an Inverted Pendulum!
Click Here! (http://youtu.be/MRuFkboBNvA)

This is a great example of programming.

[Joe G.]

Cool video!

I spy some banebots parts.

[Chadfrom308]

Cool video!

I spy some banebots parts.

I see the wheels, anything else?

I'd love to see the program on that

[Michael Hill]

I'm guessing it's a rather untuned PD controller. Inverted pendulums can be controlled pretty stably if you use even just PD. X-travel targeting can be accomplished by adding in the Integral.

We did this in my controls lab. We used LQR and PID control (though the manual only lists LQR). There's some good math behind it all though. The lab does discuss adding in an Integral controller though.

https://engineering.purdue.edu/AAE/Academics/Courses/aae364L/2011/Fall_2011/lab3/manual

These labs were always scary because we had to calculate all of our gains prior to actually testing and we had to put our gains to the test in lab (only 1 chance to get it right)

[Monochron]

Mmmm, dat PID control loop. Love this stuff.

[Chadfrom308]

Mmmm, dat PID control loop. Love this stuff.

Something tells me there is a little more than just a PID loop in there

But, the PID loop is very well tuned. Probably running insanely fast too

[mman1506]

More info here. http://www.youtube.com/watch?v=KOLRu7z-Gnk

[Taylor]

I'm guessing it's a rather untuned PD controller. Inverted pendulums can be controlled pretty stably if you use even just PD. X-travel targeting can be accomplished by adding in the Integral.

We did this in my controls lab. We used LQR and PID control (though the manual only lists LQR). There's some good math behind it all though. The lab does discuss adding in an Integral controller though.

https://engineering.purdue.edu/AAE/Academics/Courses/aae364L/2011/Fall_2011/lab3/manual

These labs were always scary because we had to calculate all of our gains prior to actually testing and we had to put our gains to the test in lab (only 1 chance to get it right)

I've seen a Boilermaker version of this machine, possibly the one Michael referenced above. However, the operator held the pendulum upright, then started the machine. How much more difficult is is for the machine to start with the pendulum pointed down, as the OP video shows?

[Monochron]

Something tells me there is a little more than just a PID loop in there

But, the PID loop is very well tuned. Probably running insanely fast too

A couple groups in my senior design class did this project a couple years ago with a PID loop and processing offloaded to XPC. Looks similar, but yeah, I would be interested to see a technical paper on how they achieved that crazy in the video :D

[Michael Hill]

I've seen a Boilermaker version of this machine, possibly the one Michael referenced above. However, the operator held the pendulum upright, then started the machine. How much more difficult is is for the machine to start with the pendulum pointed down, as the OP video shows?

It's not that much more difficult with the single pendulum to be honest. It's more difficult with the double pendulum. Part of the problem can be the physical track length. Bad things happen when you slam into the sides.

This is the version we used: http://www.quanser.com/products/linear_servo_IP02

[Tom Line]

They have a similar device at the Ann-Arbor Hands on Museum. However, the pendulum is supported from a swing - arm that travels in a circle. That way it's not on a linear track and you never run out of track to balance on.