On-Line Learning CMAC Control System

This study proposes an on-line learning CMAC (cerebellar model articulation controller) control system, which contains only one single-input controller implemented by a differentiable CMAC. The on-line learning rules derived by the gradient descent algorithm are similar to neural control schemes. However, without a preliminary off-line learning, the proposed CMAC controller can provide the control effort to an unknown plant at each time step. Simulation is given to demonstrate the effectiveness of the proposed controller.