Ricardo Santos

Evaluation of Visual Servoing Techniques for Robotic Application

Abstract

Control systems based on visual data input, termed visual servoing systems, are one of the most promising fields of research to develop a system as much autonomous as is possible.

In the last years several methods have been proposed in this area, with their strengths and weaknesses. To clarify it, it is necessary to evaluate and compare them. In this thesis some visual servoing methods are evaluated with the goal to clarify in which situations each one of them is a good option to solve a specific task. The methods evaluated are: image based visual servoing, position based visual servoing, 2.5D visual servoing, decoupled visual servoing and shortest path visual servoing.