News
- Interested students are invited to attend the information session in room 03.07.011 on Jun 23, 2016, 10:00 am
- Time and date of the course are now fixed
- The first meeting will take place on October 20
- The presentation time slots are now assigned (see table below)
Contents & Scope
Neurorobotics is an interdisciplinary field of research at the intersection of robotics, machine learning and neuroscience. It builds an experimental link between robotics and neuroscience by connecting biologically realistic simulations of the brain to robots. One the one hand, this enables neuroscientists to validate hypotheses in fully observable artificial closed-loop experiments. On the other hand, roboticist can benefit from neuroscientific findings by using brain simulations for robot control. Like real brains, these brain models cannot be programmed but must learn the desired behavior using algorithms from machine learning and neuroscience.
The focus of this practical course is twofold. First, a key element of neurorobotics research is the integration of brain simulations with robotic bodies. This includes both connecting simulators from robotics to neural modals and the interfacing of neuromorphic hardware to physical robots. The second focus of the course is on using these tools to implement and investigate different types of neurobiological learning in appropriately designed experiments. Participants will get the unique chance to realize their projects on latest state-of-the-art hardware for brain simulation.
Preliminary Topics
- Parallel power-efficient implementation of a simulator for spiking neural networks on embedded hardware
- Deep Learning with Google TensorFlow
- Autonomous driving
- Controlling a mobile robot with the SpiNNaker neuromorphic hardware
- Design of a virtual experiment for the Neurorobotics Platform
- Exploring enactive vision on a mobile robot platform
- Suggestions for further topics are highly welcome!
Required Skills
For successful completion of the lab course, proficiency in C/C++ or Python is required. Additional knowledge in robotics, embedded systems and parallel programming will be beneficial.
Final Presentation
Final presentations have been scheduled based on your preferences as listed below. The poster will be printed if you submit it before the specified date. Otherwise, you will need to take care of printing the poster yourself. The final poster has to be submitted before the presentation.
| Group |
Date |
Poster due by |
| Spiking Neural Networks for Autonomous Mobile Robots |
06.02.2017, 10:00 - 12:00 |
03.02.2017, 12:00 |
| Genetic Programming with Spiking Neural Networks on SpiNNaker |
06.02.2017, 10:00 - 12:00 |
03.02.2017, 12:00 |
| Deep Reinforcement Learning for Robotics |
10.02.2017, 10:00 - 12:00 |
09.02.2017, 12:00 |
| Deep Traffic Sign Detection on a Driving Simulator |
10.02.2017, 10:00 - 12:00 |
09.02.2017, 12:00 |
| Multilayer HTMs for Robotics |
10.02.2017, 10:00 - 12:00 |
09.02.2017, 12:00 |
Application
If you are interested in joining the lab course, please send us your personal profile. Optionally, you may also submit a project proposal. Use the templates available for download in the course materials section below. In case you submit a project as a group, please select one group member who will submit the project proposal.
The information you provide in your personal profiles and project proposals will help us to make appropriate decisions during the matching process. All applications are due by July 8, 8:00 AM (send them via e-mail to Florian Walter).
Course Materials
Further Reading
neurorobotics.net
