An Outlook on the Interplay of Artificial Intelligence and Software-Defined Metasurfaces: An Overview of Opportunities and Limitations

Abstract

Recent advances in programmable metasurfaces, also dubbed as software-defined metasurfaces (SDMs), are envisioned to offer a paradigm shift from uncontrollable to fully tunable and customizable wireless propagation environments, enabling a plethora of new applications and technological trends. Therefore, in view of this cutting edge technological concept, we first review the architecture and electromagnetic waves manipulation functionalities of SDMs. We then detail some of the recent advancements that have been made towards realizing these programmable functionalities in wireless communication applications. Furthermore, we elaborate on how artificial intelligence (AI) can address various constraints introduced by real-time deployment of SDMs, particularly in terms of latency, storage, energy efficiency, and computation. A review of the state-of-theart research on the integration of AI with SDMs is presented, highlighting their potentials as well as challenges. Finally, the paper concludes by offering a look ahead towards unexplored possibilities of AI mechanisms in the context of SDMs.

RIS-Assisted Space-Air-Ground Integrated Networks: New Horizons for Flexible Access and Connectivity

“…Hence, although RIS-aided SAGINs exhibit enhanced network reliability, it will impose an increased complexity on the network operation and optimization tasks. To address these concerns, ML is expected to be embedded in the fabric of next generation RIS-assisted SAGINs [13], rather than being a means to solve performance bottlenecks, as is currently the case with 5G. Combined with the improved computational and storage capabilities offered by space, air, and ground nodes, and the collaborative capabilities of the cloudification of the radio access network, ML will enable self-optimizing, self-organizing, and self-healing networks.…”

LiFi Through Reconfigurable Intelligent Surfaces: A New Frontier for 6G?

“…To tackle this challenge, ML-based solutions may be exploited to support the operation and optimisation of RISs, paving the way for LiFi networks to become truly autonomous, prescriptive, and predictive. Although it is envisioned that the synergy of ML and RISs will evolve the nature of LiFi applications emerging in all industries, conventional ML algorithms are not well suited to guarantee real-time user and network needs in highly dynamic and ultra low-latency-driven applications [15]. Therefore, it is essential to explore innovative mechanisms to resolve the shortcomings of existing ML approaches, such as prohibitive training and communication overhead and large processing delays, and hence, open up avenues for exciting applications across all verticals.…”

LiFi Through Reconfigurable Intelligent Surfaces: A New Frontier for 6G?

“…To tackle this challenge, ML solutions may be exploited to support RIS functions, such as maintenance, management, and operational tasks, yielding LiFi networks to be autonomous, prescriptive, and predictive. Although it is envisioned that the fusion of ML and RIS will evolve the nature of LiFi applications emerging in all industries, such as healthcare, retail, transportation, etc., conventional ML algorithms are not well suited to guarantee real-time user and network needs in highly dynamic and ultra low-latency-driven applications [14]. Therefore, it is essential to explore innovative mechanisms to resolve the shortcomings of existing ML approaches, such as prohibitive training and communication overhead and large processing delays, and hence, open up avenues for exciting applications across all verticals.…”