Pose Estimation and Position-based Visual-servoing | Stevedan Ogochukwu Omodolor Omodia

The idea of visual servoing consist of controlling the motion of a robot using information obtained from a computer vision feedback system. The visual data can come from a camera placed directly on the the robot’s end-effector(eye-in-hand configuration) or placed in a fixed position in the work space(fixed-camera configuration). The basic ideal of visual servoing consists of the minimization of an error value. The different visual-servoing techniques. differ with respect to the error function defined. In position-based image visual seroving(PBVS), the error function is specified based on position value in the robot task space. In Image-Based Visual servoing, the error function is defined in terms of image features. Although, IBVS is considered favorable compared to PBVS due to it low sensitivity to camera calibration errors, PBVS method was used. This is due to the limitation with respect to the control layer of the Tiago robot (Stevedan & Alberto, 2022).

The following shows one of the video experiments obtained during the project:

Demostration using the visual servoing technique

Some of the results of the experiments performed can be found below:

Schematic of the code design

Image after obtaining the feature point and contours

Initial configuration of the robot

References

2022

Control of the Tiago robot using Pose Estimation and Position-based visual servoing

Omodolor Stevedan, and Cortes Alberto

Jul 2022

Abs Bib PDF

In this paper, a strategy to control the position of the end-effector and the mobile base of the Tiago robot using visual feedback from a stereo camera is explained. To accomplish this, position based visual servoing technique is employed base on the estimation of position and depth information of the object been tracked. In order to ensure smooth position estimation, multple tecniques like median filtering and kalman filter was used.
@article{VisualServoing, title = {Control of the Tiago robot using Pose Estimation and Position-based visual servoing}, author = {Stevedan, Ogochukwu, Omodolor and Alberto, Sanfeliu, Cortes}, year = {2022}, my = {true} }