Research On Predictive Control And State Estimation Of Nonlinear System Output Feedback Model Based On Fuzzy Model

In industrial control systems under the influence of external interference,most of the controlled objects show strong time-varying,non-linear and other characteristics,and the system state variables are often partially or completely undetectable.Therefore,the related researches on model predictive control(MPC)design and system state estimation for nonlinear control systems have important theoretical and application values.In view of the Takagi-Sugeno(T-S)fuzzy model as an effective modeling tool for dealing with complex nonlinear control,this paper studies the output feedback model predictive control and state estimation of nonlinear control systems based on the T-S fuzzy model.In the control design and state estimation,a new strategy considering more fuzzy membership function information is proposed,and it is applied to the design of the fuzzy output feedback model predictive control controller and fuzzy state observer,effectively reducing the corresponding design.This paper obtains better control and state estimation performance than the results in the existing literature.Specifically,the research in this article is divided into the following sections:Firstly,this paper studies the design problem of fuzzy output feedback MPC under the conditions of data packet loss and unpredictable interference.It decomposes the online minimization problem of the nonlinear performance index function in the infinite time domain into two sub-problems: the design of offline fuzzy observers and the design of online relaxation output feedback MPC.More importantly,this paper uses the real-time information of fuzzy membership updated at each sampling time to propose a switching strategy with the highest priority of fuzzy membership,which significantly reduces the conservativeness of the control design.Compared with the latest results in the existing literature,the proposed method can obtain better control performance.Secondly,this paper presents a new design method of multi-time fuzzy state observer.This method is different from the traditional fuzzy state observer architecture.A new multi-time fuzzy state observer architecture that considers the current sampling time and the system information of several past sampling times is designed;Based on this,we subdivide the depth of Zhang Cheng space composed of normalized fuzzy weighting functions,and propose a novel switching mechanism based on fuzzy membership real-time ranking information to obtain multiple working modalities that can be freely switched according to real-time information.The multi-time fuzzy state observer significantly reduces the conservativeness of the discrete-time T-S fuzzy system state estimation method.In particular,this method reduces the conservativeness of the fuzzy state estimation at the same time to a certain extent at the expense of the computational cost of solving the feasible fuzzy observer.Thirdly,this paper proposes a design method of multi-time fuzzy state observer based on the idea of variable weights.This method is different from the previous method.It introduces a variable weight to develop a more effective switching mechanism based on fuzzy membership real-time ranking information,which can use the current sampling time to normalize the size of the fuzzy weighting function of different information.Furthermore,the design condition of the fuzzy observer whose conservativeness is further reduced is obtained.At the same time,the computational cost for solving feasible fuzzy observers has also been significantly reduced.That is to say,this paper obtains a more ideal result,under the premise of reducing the conservative design of the fuzzy state observer and not adding additional calculation burden.