Paper Title
Automatic Tracking of Moving Object With Eye-in-Hand Robot Manipulator

Abstract
In this paper, tracking strategies of a moving object using visual servoing are presented with an eye-in-hand robot arm. First of all, CAMshift (Continuously Adaptive Meanshift) algorithm is employed to track continuously the moving object in the image plane. Then by comparing the tracking window from CAMshift and the rectangle generated by the minimum rectangle method after back projection of the object model, a new rectangle is formed and can be seen as the object itself. Finally, object features for tracking purpose can be extracted from the new rectangle. Furthermore, through application of image Jacobian matrix, the tracking error in the image plane can be transformed to be the displacements of robot’s end effector. Accordingly, robot arm can be moved for tracking this moving object. In addition, trajectory planning of robot arm with blending technique is used for smoothly and continuously tracking the moving object. The developed object tracking strategies using eye-in-hand robot arm can be applied to an assembly task in a moving conveyor without knowing the conveyor speed or having to stop the conveyor for assembly. Keywords: Automatic Tracking, Visual Servoing, Eye-in-Hand