A Radial-Basis-Function-Network-Based Controller With Application To Controlling A Flexible Micro-Actuator

Abstract

Methods of designing a radial-basis-function-network-based (RBFN) controller and implementing it for controlling a piezopolymer bimorph flexible micro-actuator are presented. By focusing on the sigmoid function, which is generally used as the neuron activation function of neural networks and whose derivative shape is similar to that of a Gaussian function, we derive an RBFN controller by applying a differential operator to an NN controller. Applications in which a flexible micro-actuator is controlled by the RBFN controller are also described. Simulations show that the proposed controller is effective at controlling a flexible micro-actuator.