Constrained Nonlinear Systems Predictive Controller Design

Nonlinear model predictive control is to solve a nonconvex nonlinear programming controlproblem with constraints at each sampling time. NMPC attracts much attention of industrial andtheoretical researchers due to its good control performance and capability of handling constraintsexplicitly.But In fact, controlled systems are often restricted by constraints and uncertain disturbances,so this paper pays attention to how to deal with the constrained conditions to get an explicitcontrol law which has a good control performance.The main job of this paper contains two parts. Firstly, for a class of nonlinear systems withconstrained inputs an explicit fuzzy predictive control method is present. The main idea is toconstruct a terminal invariant set and explicit predictive controller with affine input on the basisof T-S fuzzy model based on fuzzy clustering algorithm. This method needn’t compute thecomplex non-convex nonlinear programming problem of earlier nonlinear predictive controlmethods and decreases the number of optimization variables and guarantees stability of theclosed-loop system. Secondly, we present a model predictive algorithm based on states observerfor a class of constrained uncertain nonlinear system with unavailable state measurements. Themain idea is to obtain an output feedback bounded controller with adjustable parameters androbust stability region by state observers and robust control Lyapunov function. In the stableregion, compute the adjustable parameter by optimizing the performance indexes to design therobust output feedback predictive controller that satisfy the constraints as well as guarantee therobust asymptotic stability of closed-loop system. Then out of the stability region we design apredictive algorithm on the basis of terminal region constraints and get the robust outputfeedback predictive controller. At last, a numerical simulation illustrates the validity of thealgorithm.