Robust Control For A Class Of Parameter Uncertain Systems

In the past three decades, the robust analysis and synthesis problem of a system family has been considerably researched. There have proposed a lot of valuable results, such as Kharitonov theorem, robust stability criterion based on quadratic stability, etc. So far, researches on robustness of interval system and linear corresponding perturbation systems are now very mature. However, researches on robustness of nonlinear corresponding perturbation systems are still not sufficient. This dissertation considers a class of basic parameter uncertainty, which is the polynomial function form of perturbation parameters. The class of uncertainty is a natural extension of the interval systems, the linear corresponding perturbation and the norm-bound one, and it is analogue to a fact hat part structured information is known. Moreover, a nonlinear function can be approximated by a polynomial, thus the class of uncertainty can be applied to research nonlinear corresponding perturbation. Furthermore, some plant models, such as flight control systems and power systems, usually are different in different operation conditions. It will result in the class of uncertainty via fitting the model parameters. Then the control problem in different operation points can be converted into a robust control design one.Motivated by the above reasons, It is significant to study the system with the class of uncertainty.Base on quadratic stability, a sufficient condition of robust stability is provided, and a method is also given to estimate the stability bound for plants with the class of uncertainties. Some existing results are special case of our ones. Methods are offered to design a stabilizer via the state and the output feedback respectively. In addition, algorithms are proposed to obtain a stabilizer such that the stability of the resulting closed loop systems is maximized with respect to the used method.Finally, an example of robust stabilization is given for some flight control systems with wide operation range, The simulation results show that the design method is in practice.