Fraunhofer Group for Microelectronics in cooperation with the Leibniz institutes IHP and FBHTechnology and hardware for neuromorphic computing
Neuromorphic computing is a new approach to designing microelectronic chips inspired by powerful and efficient biological neural networks. Neuromorphic chips can capture, analyze, and control data with ultra-low power consumption. In order to further develop the technology to meet the requirements of real applications in future products, more research is needed in terms of neuromorphic algorithms and hardware design. Non-volatile memory, in particular, is a central component of modern microelectronics and is becoming increasingly important for applications in artificial intelligence, machine learning, and neuromorphic computing.
In the EU-funded TEMPO project, 19 partners from industry and research are working on developing energy-efficient chips that will enable neuromorphic computing directly on mobile, battery-powered devices. The project aims to significantly improve the energy efficiency of neuromorphic hardware in order to allow for new applications for AI, IoT, and edge computing. The researchers are using new integrated memory technologies in innovative concepts for the implementation of analog and digital neuromorphic circuits. All exploitation levels, from applied research to IP generation and integrated systems, drive the memory and chip development. The chips designed and manufactured in the project will be used primarily for classification tasks in image recognition systems, e.g., for autonomous driving, as well as for processing other sensor data, e.g., from radar systems.
Participating FMD member institutes: Fraunhofer IPMS, EMFT und IIS