A fuzzy-neural-network-based visual feedback learning control for robot manipulators

A visual feedback learning control algorithm is proposed for a robot manipulator equipped with joint position servos employing fuzzy-membership-function-based neural networks (FMFNN), where weightings of FMFNN's are adjusted in such a way that the robot manipulator with an eye in hand is capable of not only tracking a moving object along the line of sight but also stopping in front of a static object, wherever it is. The training mechanisms of FMFNN are extended to be applied to the control of dynamic systems. To show the validity of the proposed algorithm, several numerical examples are illustrated for a robot manipulator equipped with position servos.<
>