MAD Urban
AI engineering empowering security reasoning throughout the entire lifecycle of an AI function
Start: 04.2025
End: 10.2025
The ever-increasing integration density of microchips is expanding the potential for implementing more and more comprehensive functions. However, SMEs can rarely cope with the constantly growing complexity and costs of chip design and manufacturing. Here, chiplets offer an innovative solution as they have the potential to enable application-specific microelectronics to be produced in small quantities.
The approach combines a complex system-on-a-chip (SoC) consisting of several smaller, separately manufactured functional units, known as chiplets, which communicate via a die-to-die (D2D) interconnect.
This increases the degree of reuse and allows the combination of different structural sizes. Chiplets originate in high-performance computing. Approaches for resource-constrained embedded systems are currently still largely unexplored and pose further challenges. An important aspect here is the early evaluation of the overall architecture, meaning the application-specific selection of chiplet modules and the implementation of the connection structure.
As part of CaMinoBW, the FZI Research Center for Information Technology is investigating tools and methods for designing chiplets. A central component is an integrated development and simulation environment that enables the modeling and evaluation of modular SoC architectures based on RISC-V in early design phases. The environment supports the synthesis of SoC designs for prototypical implementation on FPGAs, enabling the overall architecture to be evaluated in practice. In addition, the planned simulation environment allows targeted analysis of the interaction between the chiplets, including communication via D2D interconnects.
This research focus includes the research and design of sustainable IT innovations in the cross-sectional areas of energy, mobility, production, water management, and logistics. This involves developing systems that promote the ecologically, socially, and economically sustainable use of resources, and providing strategic consultancy services to companies, particularly SMEs, on their path to greater sustainability.
Funding notice:
The CaMinoBW joint project is funded by the Ministry of Economic Affairs, Labor and Tourism of Baden-Württemberg.
Project partners:
AI engineering empowering security reasoning throughout the entire lifecycle of an AI function
Secure and self-determined digital identities in e-commerce
Co-design for application-specific microelectronics with innovative chiplets in Baden-Württemberg
Jointly moving people and goods
Stopping pathogen transmission in hospitals quickly, reliably, and comprehensively using AI.
Generating resilient system-on-a-chip architectures for RISC-V
Body-worn sensors and wearables for recording vital data and identifying health-related issues in mental and psychosomatic disorders
Efficient and highly accurate data generation for AI applications in autonomous driving
Generative AI in Software Development with a focus on smart home applications.
Standardized interfaces for greater speed, safety, and sovereignty in vehicle development