David - anthropomorphic robot

The anthropomorphic robot David (formerly the DLR Hand Arm System) has joints with Variable Stiffness Actuators (VSA) that have mechanically adjustable flexibility in the drive train. One objective of development is to approach human capabilities, particularly with regard to dynamics, dexterity, and robustness.
The robot was first presented publicly consisting of one arm with hand  in 2010.

System description

Robotic systems are becoming ever more complex, which increases the risk of costly damages during operation. The increased risk often prevents developers from quickly testing fundamentally new regulatory and planning strategies. In addition, the dynamic characteristics of existing robot systems for dynamic, human-like motion sequences such as running or throwing are insufficient. Conventional actuators cannot provide the peak performance required for this without becoming too large and heavy. We are therefore convinced that great technological leaps in aerospace and service robotics can only be achieved with robotic systems that are robust against “everyday collisions” and have energy storage.

Mechatronics of David

David is an anthropomorphic robot developed at DLR using variable stiffness actuators (VSA). It is intended to approach its human archetype in size, weight, and performance. The focus of the development is on robustness, high dynamics, and dexterity.

Control

David is a scientific platform to experimentally investigate control methods for variable stiffness robots. The focus of the controller design includes active vibration damping, stiffness control, optimal, and cyclic motion control. Analysis and control of robotic arms and hands are treated.

Applications

The high dynamics and the robustness of the platform allow to explore new strategies for the planning and the execution. Currently we develop methods for self-awareness, in order to predict the pose and location of the robot itself using computer vision. We bring together advanced control, perception, planning, as well as manipulation skills in order to in order to solve challenging tasks with more complex interaction demands.