Inverted Pendulum Control

D

Thread Starter

danielmoulin

Hello,

I'm working on the project, which requires stabilization of a pendulum in upright position. The real system is based on servo-mechanisms and PLC, that controls the motion. First I used PID controller, but it seemed to be too slow and required very slow velocities, while moving pendulum from downright to upright position. With higher velocities, the controller didn't work at all. And just a note I'm using Bosch Rexroth PLC which works at 55 Hz. The second controller that I used was LQR-Controller, but I had problems, regarding accuracy of a mathematical model that I created - pendulum was stabilized but for very short period - about 1 to 2 sec.

Can anybody advice me how to improve the quality of control and if its even possible to control inverted pendulum at such low frequency rate? Or which method would be the most effective in this case?

Best regards
 
B

Bob Peterson

The math model ought to give you a basis to estimate the torque magnitude (and direction) required at each point in the motion. You might need to do some experimenting to see what effects friction and maybe gravity add to the mix.

I would use that as feedforward to a PID loop and just use the loop to trim it.

--
Bob
http://ilbob.blogspot.com/
 
Top