An Active Stereo Vision-Based Learning Approach for Robotic Tracking, Fixating and Grasping Control

Abstract

The following conclusions are drawn from the above experiments: (1) There exist many-to-one mapping relationships between the joint angles of the active stereo vision system and the spatial representations of the object in the workspace frame. (2) ART_NN and FF_NN can learn the mapping relationships in an invariant manner to the changing joint angles. The vision and joint angle signals of the active vision system corresponding to the object correspond to the same spatial representation of the object. 181