MPC Design for CMG based testbed

MPC Design for CMG based testbed The growing interest in space exploration is focusing the research in the development of space architectures able to be independent and self-sustainable. To this end, different control strategies have been studied for different types of spacecrafts. The following thesis has the goal to design a Model Predictive Control approach for the attitude control of a spacecraft testbed equipped with 4 Control Moment Gyroscopes in pyramidal configuration. The CMGs are torque generators used for the attitude control of spacecrafts, and the principal problem in the implementation of these devices is that, in certain conditions, is impossible to provide torque along specific a direction. This condition is called singularity, and it compromises the stability of the control system; for this reason it must be avoided. The reason why has been decided to adopt the MPC strategy is that we can directly avoid the singularity condition implementing this control approach. Contrary to other control strategies, the MPC can directly calculate the values to apply to the motors of the CMGs, without before calculate the desired value of the total output torque, avoiding the singularity condition. In the following work different simulations have been performed, investigating different MPC approaches with additional characteristics. After, the control algorithm has been converted in C++ programming language, and used to perform experiments on a testbed in pyramidal configuration located at the laboratories of Osaka University.