Robotics and Biology Laboratory

Intelligent Kinematic Problem Solving

Video of a Robot Opening a Lockbox

Bitte beachten Sie: Sobald Sie sich das Video ansehen, werden Informationen darüber an Youtube/Google übermittelt. Weitere Informationen dazu finden Sie unter Google Privacy.

Contact Persons

Manuel Baum

Oliver Brock

Project Description

The aim of this project is to investigate intelligent physical problem solving. Imagine that on its way to escape the escape room, the robot has to solve a puzzle that consist of multiple rigid bodies and joints that can lock each other, as in the figure you can see on the right side of this page. Interestingly, cockatoos can learn to solve such kinematic problems https://www.youtube.com/watch?v=rL9QOkBkbOM. But as astonishing as this feat is, behavioral biology still can not explain which characteristics empower the birds to exhibit this intelligent problem solving behavior. 

In this project, behavioral biologists and roboticists will together understand and explain the behavior in novel ways. We will identify prior knowledge and capabilities that the birds recruit for solving this task. We will model the behavior in a robotic system to identify which sensorimotor capabilities, computational principles, internal representations or prior beliefs best explain the behavior we observe in cockatoos. For this identification process we will import these priors into a robotic platform and test if they confirm our hypotheses and explain the birds’ behavior, or if we need to discard the hypotheses and consider new models. This way, we will reveal the principles underlying kinematic problem solving in general, such as representations or sensorimotor skills that biological and artificial systems, require to solve such complex kinematic problems. We will not only understand the biological behavior better, but improve our robotic systems this way too. When we identify the building blocks that enable birds to solve this problem, we can either directly include these building blocks into the robotic system, or we can extract insights that will further guide robotics research. Where necessary, we will additionally develop robotic interactive perception skills along the way, and contribute to the robotics community. In the end, we will have developed an autonomous robotic system that can solve complex kinematic puzzles.

Funding

This project is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany's Excellence Strategy - EXC 2002/1 "Science of Intelligence" - project number 390523135.

Publications

2022

Baum, Manuel; Brock, Oliver
"The World Is Its Own Best Model": Robust Real-World Manipulation Through Online Behavior Selection
Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)
2022
Baum, Manuel; Schattenhofer, Lukas; R"ossler, Theresa; Osuna-Mascaró, Antonio; Auersperg, Alice; Kacelnik, Alex; Brock, Oliver
Yoking-Based Identification of Learning Behavior in Artificial and Biological Agents
Proceedings of the International Conference on the Simulation of Adaptive Behavior 2022 -From Animals to Animats 16, Seite 67–78
Herausgeber: Springer International Publishing, Cham
2022
ISBN
978-3-031-16770-6

2021

Baum, Manuel; Brock, Oliver
Achieving Robustness in a Drawer Manipulation Task by using High-level Feedback instead of Planning
Proceedings of the DGR Days, Seite 29-29
DGR Days
2021

2017

Baum, Manuel; Brock, Oliver
Achieving Robustness by Optimizing Failure Behavior
Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), Seite 5806-5811
2017
Baum, Manuel; Bernstein, Matthew; Martín-Martín, Roberto; Höfer, Sebastian; Kulick, Johannes; Toussaint, Marc; Kacelnik, Alex; Brock, Oliver
Opening a Lockbox through Physical Exploration
Proceedings of the IEEE International Conference on Humanoid Robots (Humanoids)
2017

2015

Baum, Manuel; Meier, Martin; Schilling, Malte
Population based Mean of Multiple Computations networks: A building block for kinematic models
2015 International Joint Conference on Neural Networks (IJCNN), Seite 1–8
IEEE
2015