Research On Design Of Advanced Controller Based On Rough Set

Rough set theory is a new mathematical and artificial intelligent technique in knowledge discovery, and is becoming the basic theory in machine learning and AI. Rough set can derive control rules from process data automatically, and the control algorithm acquired by rough set can depend on process data completely, so rough set has great advantages and potential in rule based control system.This thesis does some research on building rough controller based on the problems in industry control field, and presents three renovation methods that improve the control performance of rough controller effectively.The main contents and renovation points in this thesis include the following:1. An efficient design method to improve the control performance of rough controller is presented. As the input-output data of the history process operation may not be informative enough, Extra testing signals are originally used to excite the process for acquiring sufficient data that can reflect the control laws of the operator or the existing controller. Using data from the successive exciting tests or excellent operation by operators, the rules can be updated and enriched. The effectiveness of the proposed method is demonstrated through emulating PID control, which shows that this method can improve the control performance of the rough controller and satisfying control performance can be achieved.2. The concept of simulating Bang-Bang control based on rough set is proposed by discussing the similarity between Bang-Bang control and operators' control. According to the features of Bang-Bang control, effective methods are designed to build rough controller simulating Bang-Bang control. Using uniform step signals as extra testing signals to make the rules orderliness, adding a condition attribute to make the rules layered. Simulations show that this rough controller can emulate Bang-Bang controller successfully. According to the limitation that the rough controller simulating Bang-Bang control remains steady error, an improvedmethod is proposed to combine with the rough controller simulating PID control. Simulations show that this innovated rough controller makes up the lack of the rough controller simulating Bang-Bang control and removes the steady error. 3. A design method of a parameterized controller based on rough set is presented in this paper according to the problem: When the control object is different or changed, the rough controller will have poor performance and need to be redesigned. Compared to the traditional rough controller, two tuning parameters are provided in this method to make the rough controller be applicable for different control objects. Two tuning parameters are designed to have the same effect as the proportional and integral action in PI controller. Simulations show that the parameterized rough controller can be used to different objects or different conditions.