Once you have created and designed your model predictive controller, you can review it for potential design issues. For more information, see Review Model Predictive Controller for Stability and Robustness Issues.
|Compute MPC closed-loop DC gain from output disturbances to measured outputs assuming constraints are inactive at steady state|
|Compute effect of controller tuning weights on performance|
|Size and order of MPC Controller|
|Compute steady-state value of MPC controller state for given inputs and outputs|
You can detect potential issues with your MPC controller design at the command line and using MPC Designer.
It is good practice to test the robustness of your model predictive controller to prediction errors.
Compute the closed-loop, steady-state gain for each output when a sustained, unit disturbance is added to each output.
Obtain a linear state-space model of an unconstrained MPC controller. You can use this model to analyze the frequency response and performance of the controller.
You can compare the time-domain and frequency-domain responses of multiple MPC controller designs.
Compute numerical derivatives of a closed-loop cumulated performance index with respect to weights and use them to improve model predictive controller performance.
You can analyze the optimal control sequence computed by a model predictive controller at each sample time.