Acting independently in complex and dynamic environments is natural for humans and animals. However, even modern machines and robots with learning skills are still restricted to perform special tasks in simple environments, which they are pre-programmed for. Intelligent systems, which act autonomously and which can carry out complex operations are still a challenge for research and development. These systems must be capable of completing a cycle of perception, understanding, action and learning. Therefore a complex feedback loop has to be created: The robot perceives its own motion and its environment through sensors, processes the sensor data, and computes new action commands to be executed. The outcome of these operations is being observed by the sensors again. This continuous feedback enables the robot to balance, walk, or manipulate objects, for instance.
Conventional technologies for feedback loops are often based on proportional-integral-derivative control. However, any perturbation from outside has to be explicitly classified which limits the mechanism to the use within more or less constant environments. Thus, a new learning control was needed, which enables robots to adapt and learn from experience during the operation making them more robust in order to cope with complex and unknown situations.
Stefan Schaal and his team from the Autonomous Motion Department of the Max Planck Institute in Tübingen have developed the “continuous motion optimization and control technology” which makes robots see what they manipulate. Continuously optimizing their motions and improving their “hand-eye” coordination based on a new algorithm, they can shape their behavior according to the environment and adapt to unpredictable interactions of human-robot cooperation.
Lula Robotics is now developing the technology to build high-performance computational systems and software to connect robots with the world. It is also planned that the technology will be fully integrable into existing robotic platforms. “Our system continuously optimizes its behaviour and reacts to changes giving a life-like quality that promotes close human-robot collaboration. Today, we aim to address collaborative manufacturing and assembly, but the technology might even be the basis for robots used in the home or for healthcare in the future”, explains Nathan Ratliff, co-inventor and CEO of Lula Robotics. “The team of Lula Robotics consists of scientists who have a lot of experience with the development of creating intelligent systems which are capable of optimizing their structure and features in a way that they can successfully act in a complex and changing world. Therefore, Lula Robotics is a perfect partner for the technology of the Max Planck Institute for Intelligent Systems”, says Bernd Ctortecka, Patent and Licensing manager at Max Planck Innovation.
Do you want to receive our newsletter and press releases?
Phone: (089) 290919-30