To content
Research work honored

AI-based channel models for 6G mobile networks

Preisübergabe auf dem IEEE Future Networks World Forum © CNI​/​TU Dortmund
Ko-Autor Simon Häger bei der Entgegenahme des Preises (links) und Erstautorin Melina Geis bei der Vorstellung der Arbeiten auf der 6GEM General Assembly in Dortmund (rechts)
The article, written by Melina Geis, Simon Häger, and Christian Wietfeld, deals with the prediction of mobile communications channel characteristics in newly developed frequency ranges of 26 GHz.

While current mobile networks use frequencies below 6 GHz, future industrial 6G networks will require very high data rates and will therefore be heavily reliant on higher frequencies (also known as mm waves). Their propagation characteristics are complex and sensitive to obstacles. Precise network planning therefore requires high-quality radio environmental maps, which in turn require accurate channel models. Conventional models are either too slow or not sufficiently accurate; AI-based approaches are very promising in this regard, but have not yet been applied to large production environments.

The models examined in the article capture key propagation effects and outperform empirical models with significantly lower errors and 30 times faster execution. The results show that AI models can generate fast and reliable mmWave connectivity maps, enabling efficient 6G network planning in industrial environments.

The work was carried out as part of the BMFTR-funded 6G mobile communications research at the Chair of Communication Networks at TU Dortmund University.
The Chair of Communication Networks has already received over 20 international awards in recent years.