Image Moments Based Predictive Control for Eye-in-Hand ServoingSystems

摘要

In this paper, a model predictive controller based on image moments is presented. The imagemoments are computed from a set of point features that describe a fixed object. An eye-in-handconfiguration, composed from a 6 degree of freedom manipulator robot and a video camera, isconsidered. For validation through simulation, an implementation in Matlab of the proposedarchitecture was considered. A comparison with classical proportional visual control architecturewas conducted and simulation results are revealed and commented.