Interactive Perception

We developed an RGB-D-based online algorithm for the interactive perception of articulated objects. In contrast to existing solutions to this problem, the online-nature of the algorithm permits perception during the interactions and addresses a number of shortcomings of existing methods. Our algorithm consists of three interconnected recursive estimation loops. The interplay of these loops is the key to the robustness of our proposed approach. The robustness stems from the feedback from our algorithm which can be used to adapt the robot's behavior.

Roberto Martín-Martín, Manuel Baum, Aravind Battaje, Vito Mengers

If a robot faces a novel, unseen object, it must first acquire information about the object’s kinematic structure by interacting with it. But there is an infinite number of possible ways to interact with an object. The robot therefore needs kinematic background knowledge: knowledge about the regularities that hint at the kinematic structure.

We developed a method for the efficient extraction of kinematic background knowledge from interactions with the world. We use relational model-based reinforcement learning, an approach that combines concepts from first-order logic (a relational representation) and reinforcement learning. Relational representations allow the robot to conceptualize the world as object parts and their relationship, and reinforcement learning enables it to learn from the experience it collects by interacting with the world. Using this approach, the robot is able to collect experiences and extract kinematic background knowledge that generalizes to previously unseen objects.

Sebastian Höfer, Manuel Baum, Aravind Battaje

Uncertainty is the major obstacle for robots manipulating objects in the real world. A robot can never perfectly know its position in the world, the position of objects, and the outcome of its actions. A particularly hard challenge is motion planning under uncertainty. How should the robot move, if the model of the world might be wrong or incomplete?

However, our approach reasons about uncertainty and contact. A robot can significantly reduce uncertainty if it uses contact sensing to establish controlled contact with the environment. Moreover, the robot's capabilities are increased if it anticipates contact events that can happen during the execution of the plan. Our goal is to develop algorithms that can plan under uncertainty while exploiting contact and reasoning about sensor events during planning for high dimensional motion problems.

Arne Sieverling, Előd Páll