direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Amazon Picking Challenge 2015

Bild

In May 2015, our Team RBO won a prestigious international robotics challenge, the Amazon Picking Challenge. This challenge aims to solve one of the last problems in warehouse automation: identifying and grasping objects from a warehouse shelf. Our robot was able to secure the lead by picking 10 out of 12 objects, outperforming 25 teams from Europe, USA and Asia, amongst them teams from the Massachusetts Institute of Technology (MIT), UC Berkeley as well as many robotics companies. mehr zu: Amazon Picking Challenge 2015

Physical Exploration Challenge

Bild

This project addresses a fundamental challenge in the intersection of machine learning and robotics. The machine learning community has developed formal methods to generate behaviour for agents that learn from their own actions. However, several fundamental questions are raised when trying to realize such behaviour on real-world robotics systems that shall learn to perceive, actuate and explore degrees of freedom (DoF) in the world. These questions pertain to basic theoretical aspects as well as the tight dependencies between exploration strategies and the perception and motor skills used to realize them. mehr zu: Physical Exploration Challenge

Soft Manipulation (SOMA)

Bild

The main obstacle to a wide-spread adoption of advanced manipulation systems in industry is their complexity, fragility, lack of strength, and difficulty of use. This project describes a path of disruptive innovation for the development of simple, compliant, yet strong, robust, and easy-to-program manipulation systems. The idea is: Soft Manipulation (SoMa). The project is funded by European Union's Horizon 2020 Research and Innovation Programme under Grant Agreement 645599. mehr zu: Soft Manipulation (SOMA)

Robotics-Specific Machine Learning (R-ML)

This project will develop robotics-specific machine learning methods. The requirement for such methods follows directly from the no-free-lunch theorems (Wolpert, 1996) which prove that no machine learning method works better than random guessing when averaged over all possible problems. The only way to improve over random guessing is to restrict the problem space and incorporate prior knowledge about this problem space into the learning method. mehr zu: Robotics-Specific Machine Learning (R-ML)

Photo Cross-linking/mass spectrometry (CLMS)

This project develops a structure determination method targeting proteins inaccessible by established techniques. This will be enabled by novel experimental protocols yielding high-density cross-linking and custom tailored computational methods to model protein structures from this data. mehr zu: Photo Cross-linking/mass spectrometry (CLMS)

Capabilities and consequences of recursive, hierarchical information processing in visual systems

Bild

We investigate human and robot perception. The goal is to develop a constructive understanding of perceptual information processing, capitalizing on the analytic-synthetic loop. Our implementation of this concept is based on the concept of "optical cortex and robotic interactive perception algorithms". This resemblance is so striking because it spans various levels of abstraction and matches. Indeed, this computational architecture enables predictions that match observations in humans. mehr zu: Capabilities and consequences of recursive, hierarchical information processing in visual systems

Dexterous and Sensorized Soft Robotic Hands (SCIoI - Proj. 17)

Bild

Inspired by human grasping and manipulation capabilities, we will build anthropomorphic soft robotic hands that also act as a sensor to enable robust interactions with the environment. Since they are made of soft materials, their morphology adapts to the environment which increases robustness and safety for human-robot interaction. mehr zu: Dexterous and Sensorized Soft Robotic Hands (SCIoI - Proj. 17)

Co-Design of Feedback Control and Soft Morphology for In-Hand Manipulation (SPP SMRS)

Bild

The behavior of a soft robot is determined by the robot's shape and material properties, i.e. the robot's morphology. As soft robots come into contact with their environment, they deform, implicitly performing aspects of control, sensing, and actuation. Clever morphological design therefore favorably affects the robot's behavior while at the same time simplifying control and sensing. In this project, we are co-designing the space of feedback control and morphology in the context of in-hand manipulation. Within the domain of in-hand manipulation, we develop computational tools and hardware components to support the design process, and we derive generalizable design insights that can transfer to other application of soft material robotics. mehr zu: Co-Design of Feedback Control and Soft Morphology for In-Hand Manipulation (SPP SMRS)

Parrobots

We research how birds and robots can learn to solve complex kinematic problems. To this end, we are building kinematic puzzles, called lockboxes and observe how goffin cockatoos can learn to solve these. Together with colleagues from Oxford and Vienna (where the bird experiments happen), we analyze the problem solving and learning behavior of these birds and try to extract information that may prove to be useful for a robotic approach to similar problems. mehr zu: Parrobots

Intelligent Kinematic Problem Solving

Bild

The aim of this project is to investigate intelligent physical problem solving. Our example problem is an escape room scenario, where a robot needs to find a way to escape by physically interacting with the world (e.g. pull doors open, turn keys etc) mehr zu: Intelligent Kinematic Problem Solving

Zusatzinformationen / Extras

Direktzugang:

Schnellnavigation zur Seite über Nummerneingabe