High performance computing (HPC) power in sensor systems
In the Fraunhofer-Gesellschaft's lead project NeurOSmart five institutes (ISIT, IPMS, IMS, IWU, IAIS) are jointly researching particularly energy-efficient and intelligent sensors for the next generation of autonomous systems. In this process the bridges between perception and information processing are being redefined by innovative electronics.
Intelligent machines are increasingly finding their way into our everyday lives. Obvious examples are autonomous driving or modern hoover robots at home. But even where we do not come into direct contact with them independently working robots are increasingly indispensable: as nimble and tireless helpers in logistics centers or as precise and strong partners in production environments.
The NeurOSmart flagship project focusses on the development of a new standard for intelligent hybrid computing architectures in autonomous machines and transport systems. In technological terms the sensor base - a scanning LiDAR system - will have direct access to the incoming data flows. A neuromorphic HPC chip will be coupled with an AI-supported preprocessing pipeline for data interpretation directly at the sensor. Complex communication and edge computing infrastructures are made redundant by decentralized data preprocessing and opportunities for attacking the system are reduced. The AISC-accelerator itself consists of HfO2-based FeFET crossbars and therefore has the potential to be as energy efficient as possible. Perspectively, an increase in data processing by at least two orders of magnitude is targeted.
Fraunhofer IMS leads the development of the sensor system with AI-supported data preprocessing and software-based object recognition.