Robust model predictive control based on a category of dynamic output feedback

For the polytopic uncertain systems with unmeasurable system states, this paper considers the synthesis of robust model predictive control (RMPC). A category of dynamic output feedback is adopted as the control strategy. Compared with common dynamic output feedback approach, the adopted approach adds some new freedom, which is optimized online by RMPC with some parameters of the controller given in advance. The proposed RMPC is proven to be robustly stable and recursively feasible. And the system constraints can be satisfied. Meanwhile, in order to reduce the online computational complexity of RMPC, an off-line version of the proposed output feedback RMPC is also developed. This makes the design more practical.