A brand new technical paper titled “Cut up Additive Manufacturing for Printed Neuromorphic Circuits” was printed by researchers at Karlsruher Institut für Technologie (KIT).
“Printed and versatile electronics guarantees sensible units for software domains, corresponding to sensible fast-paced client items and medical wearables, that are usually untouchable by standard inflexible silicon applied sciences. This is because of their exceptional properties corresponding to flexibility, non-toxic supplies, and having low-cost per space. Mixed with neuromorphic computing, printed neuromorphic circuits pose a lovely resolution for these software domains. Notably, the additive printing applied sciences can scale back great amount of fabrication complexities and prices. On the one hand, high-throughput additive printing processes, corresponding to roll-to-roll printing, can scale back the per-device fabrication time and price. However, jet-printing can present point-of-use customization on the expense of decrease fabrication throughput. On this work, we suggest a machine studying based mostly design framework, that respects the target and bodily constraints of break up additive manufacturing for printed neuromorphic circuits. With the proposed framework, a number of printed neural networks are educated collectively with the intention to sensibly mix a number of fabrication strategies (e.g., roll-to-roll and jet-printing). This could result in an economical fabrication of a number of completely different printed neuromorphic circuits and obtain excessive fabrication throughput, decrease price, and point-of-use customization.”
Discover the technical paper right here. Preprint printed April 2023.
Zhao, Haibin, Michael Hefenbrock, Michael Beigl, and Mehdi B. Tahoori. “Cut up Additive Manufacturing for Printed Neuromorphic Circuits.” In 2023 Design, Automation & Take a look at in Europe Convention & Exhibition (DATE). 2023.