Predictive Control For A Class Of Nonlinear Systems Based With T-S

In practical and industrial process, it appears to be difficult to describe the system within time delay and uncertain items, especially extensive complex one by utilizing accurate linear models. Our work considers the predictive controllable method which owns favourable trackability and robustness for non-linear systems. Furthermore, the comple system is constructed by utilizing the T-S fuzzy model. The contents of this paper can be separated into following four parts:1) The introduction to the background and theoretical knowledges. The objective backgrounds, the current status of recurrent predictive control together with T-S fuzzy models are demonstrated, respectively. The theoretical basis and definations for predictive controlling method and T-S fuzzy models are also introduced in detail, respectively. Besides, knowledges related with LMI and frequently-used critical lemmas are also introduced in this section.2) The problem of the Lyapunov asymptotic stability of non-linear systems is sloved based on the T-S fuzzy model. The localized controller is set up for two kinds of time-delay non-linear systems within input uncertain items and saturated actuators by utilizing the controlling law based on the feedback of the fuzzy model. Subsequently, the predictive controllable algorithm is performed and numerical simulations are further displayed to verify its feasibility.3) The finite-time predictive controlling method is studied basd on the T-S fuzzy modeled non-linear system. The following three steps are continually carried out in order to solve this problem: the"max-min"method is applied to show the up-limted value of objective functions and optimize quadratic functional indices. Subsequently, the designing method for the finite-time predictive controller is also given to ensure the status of systems bounded in a certain region in a finite period of time. This work has shown advantages and applicable values in practice. Finally, numerical simulations are also performed to verify the validity of our algorithm.4) The principle for control optimization is introduced for discrete-time non-linear system with hard restriction in the time domain and two kinds of uncertainties i.e. convex polyhedron and bounded norms. Additionally, the predictive controlling method for H∞performance is proposed. The multiple objective performance index functions simultaneously consider requirements of controlled systems and the anti-jamming capability of H∞. Furthermore, a novel sufficient condition which ensures the asymptotical stability of closed-loop systems and hard restrictions satisfied for the time domain is also proposed. Subsequently, the control law for the feedback of status is derived in analytical forms. Finally, for discrete and non-linear system with convex polyhedron and bounded norms, are further verified by numerical examples respectively. Results indicate that the designing method for the controlling system proposed in our paper has shown satisfactory controlling effects in its practical usage.