定常性能を考慮したCMAC-PID制御系の設計

In industrial processes, initiatives of production quality and cost reduction have created a demand. In such the situation, a concern with control performances in steady-state operation has been growing in recent years. On the other hand, Cerebellar Model Articulation Controller(CMAC) that is a kind of artificial neural networks has been proposed to control nonlinear systems. The CMAC has features of simple structure, a few design parameters and high-speed learning. A scheme for tuning PID parameters by using the CMACs has been also proposed. However, this scheme is focused on improvement of transient control performances. In this research, a new design of CMAC-PID control system compensating performance in steady-state is proposed. After weight tables included into the CMACs are learned, the weights that is employed in steady-state are re-tuned. Consequently, desirable control performances are obtained in both transient-state and steady-state