Dear All

Is it possible to implement the real-time model predictive control on low-cost Arduino UNO for a second order model and for horizon lengths of up to 20 steps and a 10 ms sample time ?

The state of the system must be estimated using a Lunberger observer or linear Kalman filter. (We use Arduino UNO as a PID controller for a 3KW synchronous generator excitation system.)

I've read a few articles that use model predictive control with Arduino hardware. But they mostly used Arduino as a low-level control ie. MPC calculations are done in another hardware and the results are transmit to Arduino for applying to the plant.

Regard