Abstract
PID controllers have been applied in a lot of industrial processes because its control structure is simple and physical meaning of PID parameters is clear. However the fixed PID controller cannot obtain a desired control performance because most real systems have nonlinearity. As an effective control method for these systems, a controller which has a tuning mechanism of PID parameters by using a Cerebellar Model Articulation Controller (CMAC) has been proposed (i.e, CMAC-PID controller) and the effectiveness of the method has been evaluated. However, it needs long period of time to learn CMAC in an on-line manner, therefore, it becomes a harmful effect to a practical realization of CMAC-PID controller. On the other hand, Virtual Reference Feedback Tuning (VRFT) and Fictitious Reference Iterative Tuning (FRIT) have been attracted much attention as methods which do not utilize system parameters to calculate control parameters. According to these methods, control parameters are computed by using a closed-loop data, thus it is expected to reduce a design procedure of a control system. In this paper, the new off-line learning method of CMAC unified FRIT is proposed. According to this method, CMAC learns its weights by using closed-loop data at off-line, as a result, the above practical problem of CMAC-PID controller can be resolved. Moreover the nonlinearly-expanded FRIT controller is realized by the proposed method. In this paper, the proposed method is first explained and the effectiveness of the proposed method is numerically evaluated by a simulation example.
