Vision-guided Dynamic Object Grasping of Robotic Manipulators

In order to achieve object grasping in dynamic environments for robotic manipulators, effective decisions should be made after the environment perception. When the object is moving, a moving object tracking algorithm is designed to track the object and finally realize the grasping manipulation. Based on the Robot Operating System (ROS), object grasping experiments are carried out on a 5-DOF robotic manipulator. The robotic manipulator can effectively grasp the objects on the conveyor belt, and those objects with random motion. The experimental results indicate that the moving object tracking algorithm in very effective to finish the object grasping in dynamic environments.

I. Introduction

Traditional industry field is always static structured environments where robotic manipulators often repeat the same working tasks, such as welding, assembly, transportation and so on. The fundamental manipulation of these working tasks is the object grasping in static environments [1] [2]. For the dynamic and complex industry field, robotic manipulators need to detect the real-time changes, and then change the movement to achieve the object grasping. When the object is moving, it is difficult to track the moving object. This problem is a key technical matter in the field of the object grasping during the dynamic environments [3].