Hybrid Machine Vision Control

Abstract

The tracking of objects in real time lends itself to vast industrial and practical applications; of interest is the tracking of objects within a manipulator's workspace by using machine vision. This paper presents the basics for control by vision of a 3-DOF robotic manipulator using artificial intelligence. The object or objects are first identified, next the object's movement or trajectory is estimated and then the robot's actuators are controlled to track the object of interest. The system is hierarchical with the identification process at the top of the algorithm, being carried out by a neural network (classifier) capable of identifying objects of interest and passing on the object's position within the viewing frame to the tracking process. Within the tracking process a hybrid combination of two vision processes interact to estimate the orientation and position of the object within the viewing frame, after which the hybrid process passes on appropriate information to the manipulator's control module. The manipulator control module utilizes the information from the tracking process to determine appropriate manipulator configurations in light of the ultimate tracking aim. The vision-based control is then illustrated via several experimental tests to verify its use and further improve the method.