In contrast to cloud computing with universal high-performance computing platforms that are trimmed for the highest performance, edge computing (decentralized data processing) requires highly-specialized HF, analog, mixed-signal and digital hardware designs. With this, it is a matter of achieving optimized compromises between performance and energy efficiency and tailoring this to your respective target application.
Modern signal processing and integrated circuit designs that are addressed by this program are key methods for implementing these hardware designs for the "edge". In doing so, there is a particular focus on design challenges that result from mm wave sensing and communications, as well as embedded artificial intelligence (AI) with energy-saving neuromorphic architectures.
The vision is to provide disruptive architectures of hardware/software components for smart (through to cognitive abilities) self-configuring edge devices ("things") that are designed for high performance and high bandwidth but which nevertheless consume as little energy as possible.