State Estimation Of Nonlinear System With Unknown Input

The complex nonlinear system is common in many fields,such as aerospace,communications,chemical industry and so on.The T-S fuzzy model,which was put forward by Takagi and Sugeno,is an effective method in describing this kind of system.State estimation is not only a basic aspect in control and signal processing field,but also a fundament of system reconstruction,state feedback control and fault diagnosis.In an actual control system,the unknown input tends to affect the accuracy of the state estimation,and the consequent control performance.Therefore,the study on state estimation of T-S fuzzy system with unknown input has theoretical and practical significances.On the basis of T-S fuzzy model,Lyapunov stability theory and LMI(Linear Matrix Inequality)calculation method,this paper studied the state estimation problem of T-S fuzzy system with unknown input.One of the difficulties in design unknown input filter and unknown input observer is the independence between unknown input and the H-infinity performance.This paper aims on this difficulty and the equality constraints.The processing method intends to treat the uncertain as unknown input is another innovation of this paper.The main works are as following:1.For T-S fuzzy system,two methods are put forward in unknown input observer design----Same Structure method and Model Conversion method.In order to solve the bottleneck problem in the unknown input observer design process i.e.the independence between unknown input and the H-infinity performance,the paper constructed a LMI and get the observer design conditions by constructing new system states by the original system states and observer states and defining new H_?performance.2.In addition to the unknown input,the uncertain is also considered in T-S fuzzy system with unknown input observer design problem.By proper disposal to the system model,sufficient conditions to the design of uncertain T-S fuzzy system robust unknown input observer are given by exploiting the newest decoupling technique with the uncertain treated as unknown input,which has the fewer restrictions than the existence method.3.The unbound unknown input is considered in the observer design problem.The cancellation of unknown input amplitude constraint makes not only more disturbances be considered,but also the observer design method has more range of application.Descriptor systems method is applied to the design of unknown input observer.Two cases are considered:1)the effect on the system measurement output by unknown input;2)the effect on both the system state and the system measurement output by unknown input.Under LMI technique,the sufficient conditions of unknown input observer are given.4.The paper studied the H_?controller design problem on discrete-time T-S fuzzy system based on unknown input observer.By using descriptor systems method,the paper gives the sufficient conditions of unknown input observer and unknown input observer based controller.The design conditions can ensure the closed-loop system asymptotically stable and satisfy the desired H_?performances.5.The unknown input filter design problem to T-S fuzzy system is also studied in the paper.By introducing the energy restriction and using of LMI linear transform properties,the problem that unknown input is irrelevant to H_?performances is solved successfully.This is the second method to solve the independence problem.