Design service for LQR controller for unknown system based on available Multiple Inputs and Multiple Output data

I am working on a project that require designing a MIMO temperature controller. I have very limited control experiences. Multiple inputs and multiple output data are available.

[Aquatris]

You would need to follow roughly 4 steps to get an LQR controller:

1. Obtain the FRD of the system that captures the important dynamics from the MIMO data (might require performing a new test with a carefully designed input rather than using the available data)
2. Fit a model to the captured FRD
3. Design an LQR controller using the fitted model that achieves desired performance while being robust enough for the possible modelling errors observed during step 2 via the selection of Q and R matrices
4. Test the controller on the system

Depending on your knowledge in the control area these steps might be trivial or might seem impossible. I would recommend asking more specific questions once you get stuck, for instance in the modelling the system from data or selecting the Q and R matrices etc.

Answer 1

[Sam Chak]

Hi @Hao

Do you have some free time to spare for some Prerequisite Courses so that you are ready to advance to the next level of control design? If so, consider taking this Self-Paced Online Course as the first step:

Control Design Onramp with Simulink

https://matlabacademy.mathworks.com/details/control-design-onramp-with-simulink/controls