Our groups current research activities are focused on the following three directions:
- Real-Time Systems, including formal timing analyses, scheduling algorithm design, real-time operating systems, Worst Case Execution Time, etc.
- Embedded Systems, including system software development, design and optimization for emerging architectures, etc.
- Resource-constrained Data Analyses, including error tolerant neural networks, architecture-aware data analyses, etc.
If you are interested in writing your thesis in any of the above mentioned research directions please refer to the corresponding people directly for further topic discussions or see the list of available thesis topics below.
Please see the list of completed theses to get an impression of topics provided by our group.
In case you cannot find a suitable one but you are still interested in ES/CPS designs, please refer to the aforementioned LS12 staff.
Currently available thesis topics:
- Drone Deployment and Workload Offloading in Decentralized Drone Networks (MA) !New!
- Benchmark suit for evaluating wear levelling on specific memory regions (MA/BA) !New!
- Converting and Verifying Resource Sharing Protocols in Real-Time Operating Systems (MA/BA) !New!
- Implementation and Evaluation of Stationary GANG Scheduling (BA/MA)
- Bit Error Tolerance Optimization of Object Detection Neural Networks (BA/MA)
- Architecture Search for Error Tolerant Binarized Neural Networks (BA/MA)
- Pruning Binarized Neural Networks for Embedded Systems (BA/MA)
- Dirichlet-Rescale Algorithm in Numerical Simulation on Real-Time Systems (MA)
- Hardware Acceleration for Real-time Quantum Spin Noise Spectroscopy (BA/MA)
- Energy analysis of machine learning algorithms on ultra-low power embedded devices (BA/MA)
- Hypervisor-based Memory Access Analysis and Non-Volatile Memory Simulation (BA/MA)
Ongoing thesis topics:
- Precise Access Latency Measurement for Integrated FRAM Devices (BA/MA)
- A Realistic Energy Evalution Setup for Scheudling in Multicore Embedded Non-Volatile Memory Systems (BA/MA)
- Offloading Protocols for Multi-Mode Mixed-Criticality Systems (BA)
- Non-Volatile Memory Aware Task Scheduling on Low Power Hardware (BA)
- Verifying Resource Sharing Protocols in Real-Time Operating Systems (MA)
- Efficient Inference of Random Forests on Modern Memory Architectures (BA/MA)
- Supporting Multi-processor Resource Synchronization on Erika RTOS (BA)
- Exploration of K-means Clustering Algorithms on Low-dimensional Datasets (MA)
- Verification for the Concept Drift of Hyper-Parameter when Dataset Shift (BA/MA)
- Non-Volatile Unikernels: Library Support for Non-Volatile Memory in Library OSes (MA)