Vision-Based 6-dof Robot End-effector Positioning Using Neural Networks
Huseyin H. Yakali, Ben Kröse, Leo Dorst · UvA-DARE (University of Amsterdam) · 1997
We present a method for vision-based model-free positioning of a 6-degree-of-freedom robot endeffector with respect to a planar target object using a feed-forward neural network. We investigate the necessary conditions under which a neural network can learn the mapping from feature domain to actuator domain. After satisfying these conditions, a neural network is used to learn this mapping. We consider only planar objects as target and their binary images. Moment-based image descriptors are used to represent the image in the feature domain. Simulation results are also presented. 2. Introduction The goal of vision-based robot end-effector positioning is to move the end-effector to a desired location using visuosensory information. We assume that the relative position between the endeffector and camera is fixed. Therefore, the problem becomes a camera positioning task. In order to achieve this positioning task, a controller has to be designed which maps the error in the visuosensory do...