EMSIC - Classification of the electromagnetic spectrum with machine learning methods

Project details

Description

Grid operators and organisations with security tasks must monitor the spectrum to ensure regulation-compliant use of the electromagnetic radio spectrum and prevent any risks. Wireless devices must therefore be reliably detected and classified for this purpose. As the acquisition bandwidth of modern wireless receivers is growing constantly, the volume of data to be processed is also increasing rapidly, but existing algorithms cannot be scaled for the data volumes and the increasing diversity of wireless signals. This project is therefore looking at significantly boosting the performance capability of existing radio monitoring systems. Innovative methods of machine learning will be used, with a focus on linking traditional processing methods with learning-based methods. The results will enable the use of broadband receivers in radio monitoring and therefore contribute to the safety of modern radio communication networks.