Embedded AI Localizes Data Processing for Greater Speed ​​and Security |  Research & Technology |  Nov 2022

Embedded AI Localizes Data Processing for Greater Speed ​​and Security | Research & Technology | Nov 2022

DRESDEN, Germany, Nov. 17, 2022 — The Fraunhofer Institute for Photonic Microsystems (IPMS) aims to support more secure, faster data processing by integrating machine learning algorithms into digital devices.

Although AI-enabled devices are firmly integrated with daily life, the processing of data inputs takes place on large, external servers. Embedded artificial intelligence (edge ​​AI) is poised to change this by allowing those processing tasks to take place directly on the device. However, the performance of AI, especially in very small devices, has so far been limited.


Fraunhofer IPMS is integrating artificial intelligence in microsensors and actuators. Courtesy of Fraunhofer IPMS.


The researchers at Fraunhofer IPMS are working to remedy this by networking expertise and developments from disparate research areas. For example, in an internal institute project, findings from microsensor and actuator technology were combined with the latest technologies in nanoelectronics, wireless communication, and processor developments.

The combination enables sensor- or actuator-related signal preprocessing using AI-based methods, providing advantages in low latency processing and more secure data processing while eschewing the need for network connectivity. Additionally, the use of edge AI to process data would enable re-learning locally in the field, so that the system could be optimized for specific on-site boundary conditions.

Li-Fi GigaDock from Fraunhofer IPMS enables optical data transmission of large amounts of data at low latencies.  Courtesy of Fraunhofer IPMS.


Li-Fi GigaDock from Fraunhofer IPMS enables optical data transmission of large amounts of data at low latencies. Courtesy of Fraunhofer IPMS.


Applications of the technology include spectrometers, ISFET (ion-sensitive field-effect transistor) sensors, and ultrasonic imaging for condition monitoring, gesture control, or environmental recognition for collaborative robots.

To support the integration of edge AI via sensor and actuator technologies, the researchers extended the existing RISC-V computing platform EMSA5, with AI functionality based on Tensorflow Lite. Fraunhofer demonstrated the setup at the electronica trade fair in Munich on Nov. 15-18. The researchers also presented some of the institute’s newest developments in intellectual property core technology and optical wireless data transmission at electronica.

.

Leave a Comment

Your email address will not be published. Required fields are marked *