Towards Practical Uncalibrated Visual Servoing

  • Author / Creator
    Ramirez, Oscar A
  • When controlling dynamic systems such as robots a big challenge lies in defining how the desired actions will be accomplished. In industrial settings automation has been possible due to the structured and predictable environments. The repeatability of the tasks makes it viable to manually automate them. Moving to dynamic and unstructured environments, such as human settings and the outdoors, makes such approaches impractical. Uncalibrated Visual Servoing (UVS) presents a viable approach to facilitate robot control and task definition in unstructured environments. UVS can be applied either in autonomous systems, or as a direct interface for users to manage the robot. Tasks are defined through visual features directly in image space. By estimating the full non-parametric image Jacobian no a-priori models or camera calibration is required. Real-world adoption of UVS has been slow despite over forty years of research. This work presents a minimalistic framework and accompanying software library for UVS. The goal is to enable users to create a variety of visual servoing systems using low complexity control interfaces that easily interact with visual tracking systems to produce a complete environment able to drive robot control. Our library, ROS-UVS, has now been used with several robots for a variety of tasks such as pick and place with a WAM manipulator and quadrotor assisted flight control. Our implementation, developed within the ROS framework, has proven to be flexible, robust and easy to use and integrate across multiple robots and control interfaces. A simulation environment is also available allowing users to try out our system. Finally, through the use of ROS-UVS we explore the performance of UVS both in simulation and with a physical robot. Characterizing the behaviour of UVS will help facilitate adoption for new users and serves to showcase the features and applications of our library.

  • Subjects / Keywords
  • Graduation date
    Spring 2016
  • Type of Item
  • Degree
    Master of Science
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.