Machine learning-based Wi-Fi optimisation

Wi-Fi networks are the basis for modern wireless communication. As the number of connected devices increases and new standards, such as Wi-Fi 7, are developed, there is a growing need for solutions that enable more effective management of network resources. In response to these challenges, a research team from AGH University and TU Berlin will use artificial intelligence, in particular machine learning, to optimise the performance of wireless networks.

The research activities will focus on three main areas. First, the team aims to ensure that modern devices using machine learning algorithms can interoperate with older equipment, which remains widely used in Wi-Fi networks. The researchers will investigate how the use of machine learning affects the balance between different generations of devices and how solutions can be designed to improve network performance without limiting older equipment's access to network resources.

The second area of research will explore the use of generative artificial intelligence for indoor positioning and activity detection using Wi-Fi signals. Such solutions have potential applications in smart buildings, space management systems, and location-based services. By leveraging generative AI models, researchers can reduce the amount of measurement data required while enhancing the quality of the data available for analysis.

The third task will be to improve data traffic management in Wi-Fi networks. The researchers will develop methods for more efficient transmission planning, minimising the impact of interference, and adapting network operations to changing conditions. This could prove particularly important in environments with a high density of users and devices.

The project’s results may contribute to the development of smarter and more flexible wireless networks, offering users a better quality of service and enabling new applications in homes, offices, and public spaces.

The Polish team will be led by Professor Katarzyna Kosek-Szott from the Faculty of Computer Science, Electronics, and Telecommunications. For the implementation of the project, the researcher will receive PLN 908,491 in funding from the National Science Centre. The German partner in the project is Professor Falko Dressler from the Technical University of Berlin.

***

The Weave-UNISONO call, open on an ongoing basis, is the result of multilateral cooperation between research funding organisations within Science Europe. It aims to facilitate the submission and evaluation of research projects across all scientific disciplines involving researchers from two or three European countries.

Proposals are evaluated according to the Lead Agency Procedure (LAP), which means that only one partner institution is responsible for the full substantive evaluation and the remaining partners accept the result.

As part of the Weave programme, partner research teams apply for funding to the lead agency and the respective institutions participating in the programme. The joint project must involve coherent research plans that explicitly illustrate the added value of international cooperation.