6G AI-and ML-Driven Co-Design for communication and computation: The hexa-X project's future

Abstract

Now that 5G networks are live, standards bodies are hard at work on the 6G network design phase. The complexity of 6G networks will increase the time, money, and effort needed for implementation and administration. Conversely, in order to lower operating expenditures (OPEX), mobile network operators want these networks to be smart, self-organizing, and economical. A subfield of AI known as machine learning (ML) offers practical answers to these problems and has the potential to radically alter the trajectory of wireless network technology in the future. Using a few case studies, we take a quick look at the most pressing issues, focusing on cellular networks' physical (PHY) and link (LINK) layers, where ML may make a big difference. We also take a look at the standardization efforts around ML in wireless connections and project when these organizations will be ready to adjust to new circumstances. Lastly, we draw attention to important problems with ML in wireless tech and provide suggestions for how 6G wireless networks can solve some of those problems.