Global Sampled-data Output Regulation Of Linear Systems And Low-triangular Nonlinear Systems

Output regulation, also known as servomechanism, is one of the most attractive problems in the field of automation. It involves stabilization, tracking and disturbance attenuation. The object of output regulation lies in how to design a suitable controller to asymptotically stabilize the unforced system while making the forced system track a class of signals or reject the disturbance. Currently, most of the research findings of output regulation focus on continuous controller design approach. However, with the popularity of computer control technology, how to design a discrete controller to achieve sampled-data output regulation has been of great practical value.This dissertation studies global sampled-data output regulation of both linear and a class of nonlinear systems in lower-triangular form. Based on the analysis of continuous output regulation for linear continuous systems, either the reference signals(disturbance) are constants or gengerally bounded, the linear system achieves global asympotically sampled-data output regulation by state feedback or error feedback. In order to achieve global continuous output regulation of a class of nonlinear systems in lower-triangular form, based on domination control design, a continuous state feedback controller and a continuous error feedback controller are exploited. Normally, a suitable Lyapunov function is adopted to proof the global asymptotic stability of the closed-loop system. Given the continuous results for nonlinear systems in lower-triangular form, we design a sampled-data state feedback controller as well as a sampled-data error feedback controller by discretizing the corresponding continuous controller to achieve global sampled-data output regulation of the nonlinear systems in lower-triangular form via selecting suitable sampling period. When the reference signals(disturbance) are constants, global sampled-data asymptotic output regulation is achievable, while only global sampled-data practical output regulation is achievable in the case that the reference signals are generally bounded. Simulation examples of both linear systems and nonlinear systems in lower-triangular form certify the effectiveness of the control methods proposed in the dissertation.Linear systems and nonlinear systems in lower-triangular form that are studied in the dissertation are continuous, while controllers are discrete. During the sampling period, discrete controll signals are holden to be constant by the zero-order holder. Thanks to discretization of continous controller, output regulation coupled with sampled-data controll strategy is capable of handling asymptoic output regulation of linear systems and global output regulation of nonlinear systems in lower-triangular form effectively. What is studied in the thesis can be applied to a great deal of engineering applications as theoritic foundation such as manipulator position tracking controlled by computers and motor speed regulation.