Ultra-low-power, secure processors for edge computing (RIA)

• Develop European specialised processors, architectures and computational engines that have the potential to substantially improve energy efficiency (i.e. performance per watt) for the targeted edge application.
• Examples of targeted applications (non-exhaustive list) are automated driving, artificial intelligence, machine learning, computer vision, machine translation, speech recognition, sensor fusion, signal processing, etc.
• New specialised processor designs may incorporate approaches such as neuromorphic, in-memory computing, probabilistic computing, neural networks, programmable logic, hardware-software co-design as well as open-source hardware and processor IP.
• Proposals should have a longer-term perspective taking into account the reduced performance improvements of general-purpose computing, the slow-down of Moore’s law and the changing economics of semiconductor manufacturing.
• Proposals should include research on advanced hardware-based security at silicon-level.
• Proposals should take into account certification guidelines for secure and safety-critical applications where relevant.

Proposals should include a preliminary analysis of bringing successfully to the market the proposed research either as IP blocks or as standalone chips. Proposals may include early chip prototyping in well-justified cases.

In this topic the integration of the gender dimension (sex and gender analysis) in research and innovation content is not a mandatory requirement.