The Institute of Artificial Intelligence at the University of Bremen in Germany investigates methods for cognition-enabled robot control. The research is at the intersection of robotics and Artificial Intelligence (AI) and includes methods for intelligent perception, dexterous object manipulation, plan-based robot control, and knowledge representation for robots.
Robots performing complex tasks in open domains, such as assisting humans in a household or collaboratively assembling products in a factory, need to have cognitive capabilities for interpreting their sensor data, understanding scenes, selecting and parameterizing their actions, recognizing and handling failures and interacting with humans.
In our first year in Google Summer of Code (GSoC), we have students working on three distinct projects from our core research competences:
- Mihai Baltac is working on the development of situation-specific simulation environments in the Gazebo robot simulator. On the basis of an existing plan library in the CRAM system, he will also develop robot plans that enable robotic agents to operate in this environment in a knowledge-supported, robust fashion.
- Andrei-Mihai Nicolae is improving the visualization of the belief state and the intentions of the cognitive agent. He’s further developing the Bullet physics engine based reasoning system of CRAM.
- Razvan-Andrei Stoica is extending the geometric reasoning capabilities of the KnowRob CAD reasoning system. He will introduce new attributes that can be extracted from physical properties of a known object model, and will work on further refining the current algorithms.
By Jan Winkler, Institute for Artificial Intelligence
-----------------
Frenetic is an open source software-defined networking controller platform. With Frenetic, a programmer can describe the intended behavior of the network in a high-level language, and the compiler and run-time system generates the low-level code that executes on network devices.
Software-defined networking (SDN) is an emerging network architecture in which a logically-centralized controller manages the behavior of a collection of programmable switches, such as OpenFlow switches. SDN can simplify many network algorithms, and it also makes it easy to extend the network with new functionality. Most SDN controller platforms provide low-level programming interfaces that closely mirror the capabilities of the underlying hardware.
Frenetic is unique in that it provides a high-level and declarative programming interface that abstracts away from the details of the hardware and allows programmers to focus on the essential features of network applications.
We are very excited to have a Google Summer of Code student this year. He is working hard designing and implementing support for versions 1.3 and 1.4 of the OpenFlow protocol.
By Marco Canini, Frenetic Organization Administrator