Theses

Power Management for Android-based devices

In collaboration with the UNC Chapel Hill, North Carolina, we are working on application-specific power management. There are various different aspects you could be working on, amongst others:

• Reinforcement Learning for Automatized Game Testing on Android-based Devices
• schedext - Evaluation of Application-specific Scheduling on Android-based Devices
• Development of a Testing Platform to Test the Influence of Frame Rate and Jitter on the Quality perceived by the Player

Deep-Brain Stimulation

In collaboration with the University of North Carolina at Chapel Hill, we are advancing deep brain stimulation (DBS) techniques for the treatment of Parkinson's disease. While DBS is a well-established clinical method, current devices remain relatively rudimentary. Recent research has demonstrated the significant potential of reinforcement learning to dynamically adjust stimulation frequency and amplitude in real time. However, practical application remains constrained by the limited power capacity of implantable devices. Our work focuses on developing low-power algorithms to bridge the gap between promising simulation-based results and real-world deployment in patients. Topics you could be working on:

• Power Efficient Reinforcement Learning Implementations on Implantable Devices
• Power Efficient Signal Pre-Processing of Local Field Potential Signals

Swarm-based Rapid Avalanche Search System

In collaboration with TH Rosenheim we are working on an autonomous drone based avalanche victim search system. There are various different aspects of the system you could be working on, amongst others:

• Implementation and Evaluation of an autonomous Drone-based Avalanche Search System
• Swarm based Rapid Location System for Multiple Avalanche Entrapments

Crazyflie Drones

• Web Assembly - On the fly Firmware Update for Mini-Drones
• Autonomous Wifi Mapping System

RoboSoccer Small Size League Lab

• Computer Vision Setup
• Robot Steer and Control