Hi Deepayan,
In Model Predictive Control (MPC), Measured Disturbances (MDs) are external signals that can affect the plant output and are known in advance during the control horizon. By incorporating MDs into the MPC controller design, the controller can anticipate these disturbances and adjust its actions to maintain the desired output performance.
The Plant model in MPC represents the actual system dynamics. Including MDs in the model ensures that disturbances realistically affect the plant's state and output during simulations, which mirrors real-world behaviour. This accurate representation is crucial for reliable system identification and effective real-time operation.
When MDs are included in the controller, it can predict future plant outputs more accurately. Knowing when a disturbance will occur allows the controller to proactively adjust its actions to mitigate the disturbance's impact. Although disturbances are part of the plant model, the controller needs explicit knowledge of these MDs to compute optimal control inputs effectively, as MPC relies on predicting future system states based on these disturbances.
Refer to the following MATLAB documentation link for "Model Predictive Control of Multi-Input Single-Output Plant" which describes how both plant and controller use “Measured Disturbances”:
Hope this helps.
