Design of an imitating human controller based on rough set theory

摘要

Human beings are the excellent controllers in real world. How to imitate the control-behavior of the human beings is one of the intelligent control methods. In this method, the key problem is how to model the behavior of human operator or to extract the characteristics of control-behavior of operator. The rough set theory (RST) is an important soft computing method. Some useful knowledge can be extracted from original data with it. In this paper, a design method for imitating human controller (IHC) is proposed based on RST. The computation is used to obtain the rules set (rules base) of an imitating human controller (IHC) for an actual system. The system consists of a variable frequency controller, an ac motor, a pump and a water-box. Firstly, the original data are gotten by sampling control behavior of a human operator for controlling the level of the water-box, the primal decision information table is constructed using these data, the control rules are extracted from the table according to RST, the rules set is completizated and by way of the decision rules base of the IHC, the controller is designed, and applying it to the primary system. The experimental results indicate that the IHC designed is feasible.