Position Control of Artificial Muscle


Thread Starter



I want to make a position controller in order to control an artificial fluidic muscle. I have vibration data as input (acceleration data) in m/s^2 from which I extract the vibration frequency. But in order to do the position controller, I want displacement data as input and pressure (or voltage) as output. My question is whether I can transform these acceleration data to position (double integration?) or is a better way?

Thank you

Peter Nachtwey

The short answer is yes.

Use both the acceleration and velocity as feedforward values. I have done this before. However, the acceleration data I got had a small offset that caused the position to go out of bounds over time. I had to compute and offset to apply against the acceleration data before integrating so that neither the position or velocity would go out of bounds.

This is common in testing applications.