Schedule
PLEASE CHECK BACK REGULARLY, DATES & PLACES MAY CHANGE!
What | When / How? |
---|---|
Kick Off Meeting | 19.10.2017, 15:00 - 16:30, N1039 |
Presentation Techniques | 23.10.2017, starting at 10:30, N0507 (LSR) |
Project Plan Presentation | 02.11.2017, 15:00 - 16:30, N1039 |
Project Progress Presentation | 30.11.2017, 15:00 - 16:30, N1039 |
Report Draft Submission | 14.12.2017, 12:00, per e-mail to office@nst.ei.tum.de |
Final Report Submission | 11.01.2018, 12:00, per e-mail to office@nst.ei.tum.de |
Final Presentation | 18.01.2018, 15:00 - 16:30, N1039 |
If you missed the kick-off meeting, get in touch Nicolai Waniek as soon as possible.
Topic Proposals
- Distributed Neurocomputing on Android
- Control of robotic arm with spiking neural networks
- Applying temporal coding to an existing spiking neural network for EEG motor imagery movements decoding
- A mobile app to communicate with SpiNNaker IO Board for Pushbot following laser pointer task
- Implementation of different driver AIs in The Open Source Car Racing Simulator (TORCS)
- Learning to drive based on multiple sensor cues in The Open Source Car Racing Simulator (TORCS)
- Online Decoding of Surface EMG signals for Katana Robot Arm Control
- Online Demonstration of Spike-Based EEG Decoding on SpiNNaker Neuromorphic Hardware
- Correcting Katana Robotic Arm Mistakes based on Error-Related Potentials Detection
- Optimization of an Existing Deep Convolutional Neural Network Algorithm for EEG Brain Signals Decoding
- Obstacle Avoidance with Pushbot
- Neurorobotics platform - Pushbot Race
Project Laboratory Computational Neuro Engineering
Lecturer (assistant) | |
---|---|
Number | 0000000904 |
Type | |
Duration | 4 SWS |
Term | Wintersemester 2017/18 |
Language of instruction | English |
Position within curricula | See TUMonline |
Dates | See TUMonline |
Dates
Course criteria & registration
Objectives
At the end of the module the students are able to independently design algorithms and computational hardware to process sensory information and to generate motor actions for closed-loop robotic systems. The students will understand principles of information processing in neuronal systems (as eg. distributed operations and event coding), and will be able to apply those principles to problems in technical systems. In addition, the students learn competences in efficient and problem-oriented team working.
Description
The project tasks concern hardware and software topics in the areas of neuroscientific algorithms for data processing and their application to robotics. This includes processing of high-dimensional sensory data (such as obtained from vision or audition), fusing information from such sensor modalities, and ultimately generating motor outputs that enable closed loop systems to intelligently interact in complex environments. Aspects such as hardware/software co-design, distributed parallelized information processing, limiting algorithmic complexity, and realtime data processing play an important role.
Prerequisites
Fundamentals of control engineering and robotics
Programming in C/C++ or Java
The following modules should have been successfully completed before attending this module:
Control systems 1
Fundamentals of Intelligent Robots Computational Intelligence
Examination
oral and written