Embedded Machine Learning (eMiL)

Third party funded individual grant


Acronym: eMiL

Start date : 01.10.2021

End date : 30.09.2024


Project details

Short description

This project aims to create and investigate system architectures for machine learning, dealing with several stages from embedded sensor nodes till cloud applications and in between. Their combination introduces great diversity considering energy consumption, computational performance, integration level and form factor, where workload and complexity need to be distributed through system in an optimized manner.
For demonstration, modern mmWave radar sensors are used in combination with machine learning algorithms for person presence detection and to do scans of the environment in autonomous driving situations. Investigations focus modularity, flexibility, scalability and reusability of the system.

Scientific Abstract

The aim of this project is to design and build a machine learning system that is networked across different levels, from sensors to the cloud, and optimised as a whole. The advantages of such a system can be optimally demonstrated by using the latest radar sensor technology. For this purpose, novel ML signal processing algorithms for person recognition are developed in order to realise high-resolution environment detection for autonomous transport vehicles. The focus for the system should be on modularity, reusability, flexibility and scalability, as well as the closest possible interlocking of the subcomponents.

Involved:

Contributing FAU Organisations:

Funding Source