Output Feedback Predictive Control Of Nonlinear Asymmetric Systems

Unsymmetrical system is a kind of strong nonlinear control systems. When the input directions of the control system are opposites, the steady state value and the transition process is completely different, it’s very obvious around the equilibrium point. So the control performance of the controllers razed on the system’s linear model is not ideal, and the control quality can’t be guaranteed. Based on the model predictive control, unsymmetrical process predictive control is a control algorithm. It uses the advantages of the online rolling optimization and feedback correction about model predictive control, to overcome the factors of modeling errors, external disturbances and parameter perturbation about unsymmetrical controlled object.Then through the comparison of the controller performance index of the positive and negative controllers, it determines the final control action and achieves the expected goal. Model predictive control(MPC)[1] not only can solve the state constraints of the system’s input and output, but also can control the system well. Therefore, it is widely used in the actual industrial process, and it has been greatly extended and researched in the academic domain.The cardinal contents of this article include the following two parts:Firstly, an output feedback predictive controller is designed for a class of unsymmetrical systems with immeasurable states. Since the control effect of the traditional predictive control method is not good, and output feedback is easier to achieve compared with state feedback in practice. Furthermore, in this paper the unsymmetrical system model is modified through the output feedback predictive control method. The ultimate control action was determined by the comparison of control performance index of the positive/negative controllers. And the actual control action that applied to the object is joined into the calculation of the next step control law of the controllers through the feedback form in order to eliminate the effect.Secondly, an output feedback predictive controller is designed for a kind of time-delay nonlinear systems by using the receding optimization of the predictive control. The complementary linearization idea is then employed to convert the controller design into a nonlinear programming problem with linear matrix inequality constrains, and a construction of the desired controllers is also provided in terms of feasible solutions to linear matrix inequality. The ultimate control action wasdetermined by the comparison of control performance index of the positive/negative controllers. So as to ensure the system to stabilize with the fastest speed. Finally, the effectiveness of the proposed method is verified by simulation.