Visual Error Constraint Free Visual Servoing using Novel Switched Part Jacobian Control

Visual Error Constraint Free Visual Servoing using Novel Switched Part Jacobian Control

This study extends the state of image-based visual servoing(IBVS) to handle large visual error constraints in the joint space of a robotic arm without the camera advance/retreat problem. Conventional control schemes comprise either a partitioned or a switched system, utilizing the feature Jacobian a...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access Vol. 10; p. 1
Main Authors: Arif, Zubair, Fu, Yili, Siddiqui, Muhammad Kashif, Zhang, Fuhai
Format: Journal Article
Language:English
Published: Piscataway IEEE 2022
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Camera advance / Retreat motion
Camera advance/retreat problem
Cameras
Computational efficiency
Controllers
Convergence
Degrees of freedom
Errors
Image-based-visual servoing (IBVS)
Jacobi matrix method
Jacobian matrices
Jacobian matrix
large orientation error
Manipulator Jacobians
Optical imaging
Partitioned Jacobian
Robot arms
Robotics
Robots
Switching Control
vision-based control
Visual control
Visual servoing
Visualization
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study extends the state of image-based visual servoing(IBVS) to handle large visual error constraints in the joint space of a robotic arm without the camera advance/retreat problem. Conventional control schemes comprise either a partitioned or a switched system, utilizing the feature Jacobian and rely on finding the specific feature for partitioning along the specific DOF. We proposed a novel part manipulator Jacobian method for devising a hybrid-switched-part Jacobian-based IBVS scheme, without the need to define new features. This novel and computationally efficient direct user-defined part end effector Jacobian was utilized to build an adaptive gain switching control scheme. We validated the proposed scheme for an eye-in-hand configuration on a 6 -DoF robotic arm in simulation and, on a 7-DoF real robotic arm for a set of large visual errors including 180° rotational error. The proposed scheme was compared with available IBVS methods where the proposed scheme demonstrated better performance than the Jacobian pseudo-inverse and the second-order minimization-based IBVS schemes. The experimental results showed a significant increase in the convergence zone for rotational errors up to 180° with a 40% increase in the convergence rates along with a remarkable 90% reduction in the joint velocities and energy required to complete the task. The proposed controller carries no camera advance/retreat motion, is computationally efficient, and has a well-conditioned task Jacobian matrix. In addition, the method is invariant to the robot DoF and can be extended to other visual servoing schemes.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2022.3203734