Single-Input CMAC Control System with Direct Control Ability

Without the help of any workable controller, a single-input cerebellar model articulation controller (CMAC) control system, which contains only one controller implemented by the one-dimensional CMAC, is proposed in this study. Without requiring a preliminary off-line learning phase, the single-input CMAC controller can still provide the control effort to the plant at each time step. In addition, the proposed controller is also in the learning stage. Hence the single-input CMAC controller is said to have the direct control ability. This study also introduces the adaptive deadzone size into the learning rule to solve the instability caused by the continued learning of the CMAC. Simulations on controlling a nonlinear discrete plant are given to demonstrate the effectiveness of the proposed controller.