Light - Weight Robots


The DLR Light Weight Robot is used in two different research approaches. One is the use for research on human grasping in combination with the DLR artificial Hand II. The other is the use for technical service tasks with different tools or grippers. In most cases the robot is one way or other sensor guided to achieve its goal.

In the field of grasp research the robot is used to demonstrate service tasks with varying objects, both stationary or on a mobile platform.

Exemplary applications are:

  • Picking and handling objects from a desk or shelf, which are determined by stereoscopic cameras (Robutler)
  • Catching a soft ball thrown by a person and tracked by stereoscopic cameras
  • Grasping objects telemanipulated by a data glove and moving them by a Polhemus Tracker as shown in the Knoff-Hoff-Show on German TV

In the field of tool based service robotics the robot fulfils a task using internal and/or external sensors for guidance or obstacle avoidance.

Exemplary applications are:

  • Assembling processes narrower than the system’s accuracy
  • Handling an inverse pendulum and by the time giving way when external influences occur
  • Trailing and determining an unknown contour by keeping up a contact force

Pistion Insertion

Teaching by demonstration is a typical application for the impedance controller structure. A practical example was given with the task teaching and automatic insertion of a piston into a motor block. Teaching is realized by guiding the robot with the human hand, just using the internal torque sensing.

It was initially known that the axes of the holes in the motor block were vertically oriented. In the teaching phase, high stiffness components for the orientations were commanded, while the translational stiffness was set to zero. This allowed only translational movements to be demonstrated by the human operator.

In the second phase, the taught trajectory has been automatically reproduced by the robot. In this phase, high values were assigned for the translational stiffness, while the stiffness for the rotations was low. This enabled the robot to compensate for the remaining position errors. In this experiment, the assembly was executed automatically four times faster than by the human operator in the teaching phase. For two pistons, the total time for the assembly was 6s. The insertion task has been implemented before using an industrial robot and a compliant force-torque sensor. Despite a well tuned Cartesian force controller, the insertion process had to be performed much slower, because of the well known control problems which occur in case of hard contacts with conventional robots. Thus, the advantage of a compliant manipulator became clear.

URL for this article
Links zu diesem Artikel