Several Novel Methods Of Variable Universe Adaptive Fuzzy Control

Variable universe adaptive fuzzy control is a type of adaptive fuzzy control with dynamic adjustment domain. Compared with the traditional adaptive fuzzy control, it has many ex-cellent properties, such as fast response, high steady precision, and so on. Taking variable universe adaptive fuzzy control as research background, this dissertation focuses on some prob-lems appeared in the existing literatures, such as parameter design of contraction-expansion factor, dependency on the initial control rule base, constraints on system uncertainty and fuzzy approximation error, immeasurable state, and so on. Several novel methods for solving these problems are proposed. The research work in this dissertation mainly includes following parts:1. Through the analysis of the impact of contraction-expansion factor parameters on the transient and steady state performance of control systems, some design strategies for these parameters are derived. Firstly, a stretching function is defined and the relation between control law, the stretching function and control performance is analyzed. In the transient state, the higher the derivative of control law to parameter is, the faster the system response is. In the steady state, the bigger the slope of the stretching function is, the more sensitive the system to the change of input variable is. Then, through the study of the derivative of control quantity to the parameters and the slope of the stretching function near the equilibrium point, the effect of the parameters on the transient and steady state performance of the control system are analyzed and the design principle of each parameter is obtained.2. In this dissertation, using Lyapunov method, three variable universe adaptive fuzzy con-trol schemes are put forward. Firstly, a symbolic function is employed to replace the stretching operation of input variable, and a symbol-type adaptive fuzzy control scheme is provided. The scheme not only ensures the real-time capability and stability of the closed-loop system, but also avoids the design of contraction-expansion factor. Secondly, the original parameter vector and the integral regulation factor are blended together to form a new parameter vector, and an adjusted parameter vector-based variable universe adaptive fuzzy control is proposed. The control scheme can be less dependent on the initial control rules. The third one is a robust variable universe adaptive fuzzy control scheme. By using adaptive boundary technique and σ-modification technique, the scheme can not only remove the constraints on nonlinear function and fuzzy approxima-tion error, but also guarantee the stability of the closed-loop system.3. For a type of Multiple Input Multiple Output (MIMO) nonlinear system with uncertain-ties and external disturbances, two direct adaptive fuzzy sliding mode control schemes are developed. In the first one, two fuzzy systems are used to estimate the equivalent control law and the switching control law. In the second one, a fuzzy system is used to estimate the equivalent control law, and a variable universe fuzzy system is used to approximate the switching control law. In order to remove the chattering phenomena of control signal and speed the system response, fuzzy switching control terms (in scheme one) or variable universe fuzzy switching control terms (in scheme two) are utilized. And compensation control terms are applied to reduce the influence of fuzzy approximation error and exter-nal disturbances. Based on Lyapnuov stability theory, both of these schemes can ensure the stability of the closed-loop system and achieve the tracking task. In the presence of system uncertainties and external disturbances, using these two proposed schemes, the equivalent control law can be effectively calculated; also, the chattering phenomena of the control signal can be avoided.4. For a type of Single Input Single Output (SISO) system with unmeasurable state variables, an observer-based direct adaptive fuzzy control scheme is developed. Firstly, a modified ideal control law is built. Then, a variable universe fuzzy system is employed to approx-imate the modified ideal control law. Finally, a robust control term and a tracking error estimation feedback control term are designed to suppress the influence of the lumped un-certainties and reduce the observation error, respectively. Based on SPR-Lyapunov design method, the proposed control scheme not only guarantees the stability of the closed-loop system under the condition of immeasurable states, but also requires no additional dy-namic orders for the control design.