Abstract
A simple and effective method for camera positioning and alignment control for robotic pick-and-place tasks is described here. A parallelogram feature is encoded into each 3D object or target location. To determine the pose of each part and guide the robot precisely, a camera is mounted on the robot end-flange to determine and measure the location and pose of the part. The robot then adjusts the camera to align it with the located part so that it can be grasped. The overall robot control system follows a continuous look-and-move control strategy. After a position-based coarse alignment, a sequence of image-based fine alignment steps is carried out, and the part is then picked and placed by the robot arm gripper. Experimental results showed an excellent applicability of the proposed approach for pick-and-place tasks, and the overall errors were 1.2 mm for positioning and 1.0° for orientation angle.
Highlights
There is an abundance of different literature on vision-based robot control
This paper describes a simple and efficient method for camera positioning and alignment control using a single eye-in-hand camera
The main advantages of the proposed visual servoing are as follows: (1) easy implementation and less calibration; (2) a planar image feature based on parallelism without the need to compute the image Jacobian matrix; and (3) the ability to measure precise absolute depth information by a single eye-in-hand camera
Summary
There is an abundance of different literature on vision-based robot control. In this technology, image information is used to measure the error between the current position and orientation (pose) of a robot and its desired reference pose [1,2]. Another study presented a fast peg-and-hole alignment system, in which a visual servoing approach with a single eye-in-hand high-speed camera is used to overcome the problem of position and attitude uncertainty [12]. The primary objective of this paper is the study of vision-guided robotic pick-and-place control systems with only one eye-in-hand camera based on a simple parallelogram model. A simple and effective method for camera positioning and alignment control during robotic pick-and-place tasks is offered.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have