As 5G networks continue to roll out globally and the vision for 6G begins to take shape, the demand for precise and reliable location data has never been greater. Applications ranging from autonomous vehicles and smart cities to industrial automation and augmented reality rely on robust positioning capabilities.
M.Sc. Roman Klus’s work proposes a scalable and intelligent solution that leverages deep neural networks to interpret signal characteristics and environmental context, enabling devices to localize themselves with high accuracy — even in non-line-of-sight and other challenging conditions.
Location-awareness is more than just localization
In the past, location services were used solely for localization, either for navigation or entertainment. In contrast, location-awareness encompasses a broader concept where the communication system not only knows the location but also uses that information to adapt its behaviour, optimize performance, and enable context-aware services. Location-awareness integrates localization data into network functions, enabling intelligent decision-making for tasks like resource allocation, mobility management, and service personalization.
“Utilizing AI models at both ends of the network enables autonomous, optimized, and self-learning operation of the communication systems leading to improved spectral efficiency, latency, and radio link reliability,” Roman Klus.
“The accurate knowledge of the user’s current position can be exploited to support these due to the necessity of beamforming-based communication at millimeter-wave bands,” he continues.
Integration of deep learning within 5G and beyond
The dissertation presents a various architectures utilizing e.g. convolutional and recurrent neural networks, optimized for real-time inference and adaptable to various deployment scenarios. It also explores the integration of location-awareness into the broader network stack, offering insights into how future wireless systems can become AI-driven and context-aware, while seamlessly adapting to the changes in the environment.
The findings have already attracted interest from industry partners working on 6G technologies. Klus’s work contributes to the growing body of research that positions Finland as a leader in wireless innovation and AI-driven network design.
Klus conducted his research during 2019–2025 at Tampere University, Tampere, Finland. During this period, he conducted a research stay at Universidad Jaume I., Castellón de la Plana (UJI), Spain, in 2022 and at University of California, Los Angeles (UCLA), USA in 2024. He also participated in numerous industrial projects with the leading companies in the field.
Public defence on Friday 5 September
The doctoral dissertation of Ing. Roman Klus in the field of communications engineering titled Deep Learning for Seamless Location-Awareness in 5G and Beyond Wireless Networks will be publicly examined at the Faculty of Information Technology and Communication Sciences at Tampere University at 12:00 noon on Friday 05.09.2025 at Hervanta campus, Tietotalo, auditorium TB109 (Korkeakoulunkatu 1, Tampere).
The Opponent will be Prof. Risto Wichman from Aalto University. The Custos will be Prof. Mikko Valkama from Faculty of Information Technology and Communication Sciences at Tampere University. The work has been co-supervised by D.Sc. (Tech) Jukka Talvitie from Tampere University.
The doctoral dissertation is available online.
The public defence can be followed via remote connection.