Image-based visual servoing for dual-arm manipulation in the PR2 robot


Robot Control, Vision, Programming

The images depict the implementation of the visual servoing control implemented in a industrial robot and the PR2 robot.



The goal of this thesis is to implement a uncalibrated image-based visual servoing in the PR2 robot to control the arms. The PR2 has two 7-DOF arms and a pan-tilt head equiped with 2 cameras and a Kinect sensor. 

Tasks and Objectives

The task of this project will be:

  1. Implement an AR marker tracker in ROS (C++), using ArUco.
  2. Implement the visual-servoing control for each arm using ROS (C++) see: Uncalibrated 3D Stereo Image-based Dynamic Visual Servoing for Robot Manipulators
  3. Design a visual-based dual-arm task to validate the control algorithm.  


  • Basic robot knowledge
  • C/C++ programming skills (Linux)