Embedded Systems
The Chair for Embedded Systems addresses design, analysis, and verification of distributed embedded systems and systems-on-chip as well as their application in the automotive, avionics, internet of things (IoT) and the medical domain. The trend towards autonomous systems in various application domains accompanied with the increasing demand for safety and security enforces the need for efficient stress tests under varying environmental conditions in order to build robust application-specific hardware/software architectures especially in the case of implementing machine-learning applications on embedded devices. Our research activities in the areas of embedded AI architectures and robust perception is taking place a.o. in cooperation with the Cyber Valley initiative.
Embedded software systems
- Analysis and optimization of performance, power and energy demand
- Generation of timing/power predictable software
- Hardware/software evaluation for technical feasibility assessment
Architectural design – from system-level to tapeout
- Energy-efficient AI architectures (“Edge AI”)
- Architectural exploration & technology projection
- Hardware-enhanced security architectures
AI system hardware/software co-design
- Generation of application-specific ML accelerators
- Resource and energy efficient machine learning
- Hardware-specific optimization of artificial neural networks
Robust perception in cyber-physical systems
- Modeling and simulation of environmental conditions
- Sensor (video, radar) and communication models
- Robust learning with respect to varying environmental conditions