Stable adaptive model predictive control for nonlinear systems

This research investigates the development of a stable adaptive model predictive control approach for a constrained nonlinear system. The method is well- known as multistep Newton-type control strategies however, the formulation here differs from the original one. The nonlinear physical equations of the system are extracted considering all possible effective forces. The nonlinear model is adaptively linearized during prediction procedure. The linearization not only takes place at each sampling instant of the control system, but also at each instant of the prediction horizon. The first step of this research is devoted to develop linearized models in the operating points, which are unknown and desired. Developing the equations to form a linear quadratic objective function with constraints is then carried out. Finally, the stability of the control system is provided using terminal equality constraints. To show the effectiveness of the proposed method, it is applied on a constrained highly nonlinear aerodynamic test bed, twin rotor MIMO system (TRMS).