Projects

Phase 2 Projects

Learning modular policies for robot motor skills


Project leaders: Gerhard Neumann (TU Darmstadt), Jan Peters (TU Darmstadt)

Researchers: Christian Daniel (TU Darmstadt), Sebastian Gomez (TU Darmstadt), Riad Akrour (TU Darmstadt)

Administration: Veronika Weber (TU Darmstadt)

Associates:

Summary:

The goal of this project is to develop a autonomous learning system that enables robots to autonomously acquire and improve a rich set of motor skills. The basic concept of our autonomous learning system is to decompose complex motor skills into simpler elemental movements, also called movement primitives, that serve as building blocks of our movement strategy. For example, in a tennis game, such primitives can represent different tennis strokes such as a forehand stroke, a backhand stroke or a smash. For walking, different primitives can be used for foot placement, while an additional primitive takes care of maintaining balance by shifting the center of mass of the robot. As we can see, the autonomous decomposition into building blocks is inherent to many motor tasks. In this project, we want to exploit this basic structure for our learning system.

To do so, our autonomous learning system has to extract the movement primitives out of observed trajectories, learn to generalize the primitives to different situations and select between, sequence or combine the movement primitives such that complex behavior can be synthesized out of the primitive building blocks. Our autonomous learning system will be applicable to learning from demonstrations as well as subsequent self improvement
by reinforcement learning. Learning will take place on several layers of the modular policy. While on the upper level, we will learn the activation policy for the building blocks, the intermediate level extracts metaparameters of the primitives and autonomously learns how to adapt these parameters to the current situation. At the lowest level, we have to learn the control policies of the single primitives. Learning on all layers as well as the extraction of the structure of the modular policy is aimed to operate with a minimal amount of dependence from a human expert. We will evaluate our autonomous learning framework on a robot table tennis platform, which will give us many insights in the modular structure of complex motor tasks.

Phase 1 Projects

Learning modular policies for robot motor skills


Project leaders: Gerhard Neumann (TU Darmstadt), Jan Peters (TU Darmstadt)

Researchers: Christian Daniel (TU Darmstadt), Sebastian Gomez (TU Darmstadt), Riad Akrour (TU Darmstadt)

Administration: Veronika Weber (TU Darmstadt)

Associates:

Summary:

The goal of this project is to develop a autonomous learning system that enables robots to autonomously acquire and improve a rich set of motor skills. The basic concept of our autonomous learning system is to decompose complex motor skills into simpler elemental movements, also called movement primitives, that serve as building blocks of our movement strategy. For example, in a tennis game, such primitives can represent different tennis strokes such as a forehand stroke, a backhand stroke or a smash. For walking, different primitives can be used for foot placement, while an additional primitive takes care of maintaining balance by shifting the center of mass of the robot. As we can see, the autonomous decomposition into building blocks is inherent to many motor tasks. In this project, we want to exploit this basic structure for our learning system.

To do so, our autonomous learning system has to extract the movement primitives out of observed trajectories, learn to generalize the primitives to different situations and select between, sequence or combine the movement primitives such that complex behavior can be synthesized out of the primitive building blocks. Our autonomous learning system will be applicable to learning from demonstrations as well as subsequent self improvement
by reinforcement learning. Learning will take place on several layers of the modular policy. While on the upper level, we will learn the activation policy for the building blocks, the intermediate level extracts metaparameters of the primitives and autonomously learns how to adapt these parameters to the current situation. At the lowest level, we have to learn the control policies of the single primitives. Learning on all layers as well as the extraction of the structure of the modular policy is aimed to operate with a minimal amount of dependence from a human expert. We will evaluate our autonomous learning framework on a robot table tennis platform, which will give us many insights in the modular structure of complex motor tasks.