Visual servoing of UR robots

Human Robotics has recently acquired an UR3 from Universal Robots. The researchers of Human Robotics have integrated both the robot and vision systems in ROS to perform visual servoing tasks. This kind of robot controllers allow to perform robot guidance by using visual information without determining the 3D pose of the robot workspace. The control is directly performed in the image space like human beings do. Therefore, the implemented image-based visual servoing system allows to increase the performance of robots, performing its guidance with respect to observed object in the workspace. Different visual servoing control laws have been tested to perform the tracking of trajectories and the position of the robot in a robust way. The implementation has been integrated in ROS which allows to increase the flexibility of the controllers and the easy integration of new sensor systems and to extent the controller to new robots. This kind of controllers can be applied to different social and industrial environments like humanoid robots interacting with humans or grasping objects in unstructured environments.