AI-assisted PHY technologies for 6G and beyond wireless networks
R. Sattiraju, A. Weinand, HD. Schotten. arXiv preprint 2019
[Arxiv]
Machine Learning (ML) and Artificial Intelligence (AI) have become alternative approaches in wireless networks beside conventional approaches such as model based solution concepts. Whereas traditional design concepts include the modelling of the behaviour of the underlying processes, AI based approaches allow to design network functions by learning from input data which is supposed to get mapped to specific outputs (training). Additionally, new input/output relations can be learnt during the deployement phase of the function (online learning) and make AI based solutions flexible, in order to react to new situations. Especially, new introduced use cases such as Ultra Reliable Low Latency Communication (URLLC) and Massive Machine Type Communications (MMTC) in 5G make this approach necessary, as the network complexity is further enhanced compared to networks mainly designed for human driven traffic (4G, 5G xMBB). The focus of this paper is to illustrate exemplary applications of AI techniques at the Physical Layer (PHY) of future wireless systems and therfore they can be seen as candidate technologies for e.g. 6G systems.