Does my PI + MPC (feedforward controller) configuration make sense?

Hi,

I have designed a feedforward controller to improve the tracking performance of a device. That work has been done already my purpose is to first evaluate my controller on the same work and then I will change the method to make my contribution. Here is a paper link (https://www.sciencedirect.com/science/article/pii/S0957415818301831)

But I came to know that I have some limitations and now I am consulting here for guidance. The paper I am following has used two sensors and it makes sense, but I have only one sensor. The block diagram with the red pen shows my configuration and the black one shows the other one I am following.

The difference is that I am feeding the same signal for both controllers but the following papers fed the piezo actuator signal and tip-sample signal separately as y1, and y2. The final goal is to improve the tracking performance of the piezo actuator and increase its tracking speed by compensating for nonlinearities like creep and hysteresis.

Answer 1

[Emmanouil Tzorakoleftherakis]

Looks like the whole point of using an MPC controller was to provide deltas on the PI output based on the output of the piezo actuator. If you don't have access to it, there is not much point in using the MPC loop. If you have experimental data, you could potentially create a 'virtual sensor' that estimates the output of the piezo actuator. That way you can follow the architecture in black.