Repetitive robot visual servoing via segmented gained neural network controller

Abstract

The purpose of this paper is to design a neural network controller for a nonlinear system with uncertainties which are invariant or repetitive over repeatedly executed tasks such that the maximum tracking errors can be kept within a predefined region through an iterative learning or training process. The desired trajectory is segmented and for each segment a local neural network is constructed. The training of the local neural networks is done iteratively as the task repeats. Meanwhile, the training is segment-wise progressed from the starting segment to the ending one. The accurate tracking of the whole desired trajectory is thus accomplished in a step-by-step or segment-by-segment manner. As an application example, a robot visual servoing control problem is considered with an unknown system structure and camera parameters.