6G-AINA - 6G AI-native Air Interface

The 6G-AINA project aims to contribute algorithmic innovations towards the realization of a truly AI-native air interface for 6G communication systems.

In this context, we have identified several key enabling technologies for next-generation radio systems, namely distributed computation over the wireless medium, AI/ML algorithms such as multi-agent deep reinforcement learning, as well as software-defined localization and positioning. The deepening of our understanding and the tailoring of these tools for advancing high-performing communication systems, is the centerpiece of the 6G-AINA research roadmap. In particular, we will leverage AI/ML algorithms to improve multi-antenna rate-splitting and non-orthogonal (MIMO-RSMA/NOMA) multiple-access schemes, extend current proposals of deep learning-based coded modulation and waveform designs, as well as to enhance the accuracy, sustainability and bandwidth efficiency of extremely large antenna arrays and cell-free massive MIMO networks.

The project will combine and apply these techniques to build prototypes of AI-native radio system components and validate their beyond state-of-the-art performance by proof-of-concept implementations on an experimental platform.

For the experimental part, the project counts with modern laboratory equipment at its fingertips. This includes powerful GPU-equipped workstations and software-defined radio modules for over-the-air validation of AI/ML physical-layer algorithms, as well as, for localization and synchronization, a very mature testbed for software-defined radio GNSS positioning (GNSS-SDR). We pursue an ambitious internationalization plan which revolves around our participation in competitive programmes such as Horizon Europe projects (6G-SNS and Cluster 4 Workprogrames, mostly), European Space Agency, etc. We also plan to exploit selected project outcomes via R&D industrial contracts and, eventually, raise the Technology Readiness Level to favour their exploitation.