I am trying to control wheeled inverted pendulum using a LQR is there a technique by trial and error to tuning a K constants for (X Xdot Teta Tetadot)? like pid controller we have to begin by increasing a proportional error until we get a first vibration and ....
No,
in this case, it is not LQR anymore. it would be simple state feedback controller and if you could tune it by trail and error, it would not work appropriately.
first try to stabilize the pendulum around 0 by feedbacking teta and tetadot by finding proper gains. Then tune the x and xdot gains.
or
employ pole placement approach to find the gains.
you can see:
Katsuhiko Ogata, Modern Control Engineering, Prentice Hall, 2010 .
Hi
take a look at the following link
https://www.researchgate.net/publication/274071644_On_the_design_of_an_autopilot_for_roll_motion_in_an_airplane_using_LQR_and_flexible_dynamic_model
Conference Paper On the design of an autopilot for roll motion in an airplane...
The weight in LQG can be tuned (with a proven convergent algorithm by [Zhu,Skelton, Rotea]) to solve the following problem. Find the minimal control energy that can guarantee a prespecified upper bound on the covariance of the output error.
I believe that the paper is in the 1980s, but i am on travel and cannot send it to you.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory