TEMPO - Technology and Hardware for Neuromorphic Computing
Project duration: 2019-2022
Neuromorphic computing is a new approach to microelectronic chip design inspired by the powerful and efficient biological neural networks. Neuromorphic chips are capable of acquiring, analyzing, and controlling data with ultra-low power consumption. In order to further develop the technology to meet the requirements of real-world applications in future products, more research is needed in terms of neuromorphic algorithms and hardware design.
The TEMPO project therefore aims to significantly improve energy efficiency of neuromorphic hardware to enable novel applications in industry, healthcare, assistance systems and many more.
Fraunhofer IPMS is involved in the technological implementation of microelectronic components with particularly low power consumption, which will also be usable for applications in the field of artificial intelligence, IoT and edge computing. One partial aspect is novel memory concepts as well as the integration of additional functional modules made of ferroelectric capacitors into the BEoL level of chips for the realization of BEoL-FeFET memory arrays. In addition, Fraunhofer IPMS is cooperating directly with industrial partners within the project in order to be able to ensure the subsequent industrial feasibility of the neuromorphic circuits and chips.