Design and validation of an MPC controller for CMG-based testbed

In the last years, Control Moment Gyros (CMGs) are widely used for high-speed attitude control, since they are able to generate larger torque compared to “classical” actuation systems, such as Reaction Wheels . This paper describes the attitude control problem of a spacecraft, using a Model Predictive Control method. The features of the considered linear MPC are: (i) a virtual reference, to guarantee input constraints satisfaction, and (ii) an integrator state as a servo compensator, to reduce the steady-state error. Moreover, the real-time implementability is investigated using an experimental testbed with four CMGs in pyramidal configuration, where the capability of attitude control and the optimization solver for embedded systems are focused on. The effectiveness and the performance of the control system are shown in both simulations and experiments.