A Treatise on the Taxonomy of Reconfigurable Intelligent Surfaces

This work presents a comprehensive taxonomy of reconfigurable intelligent surfaces (RISs) across its many variants in a systematically-structured form. The taxonomy encapsulates the different RIS categories based on the RIS components, channel estimation capability, size, mobility, shape, elements’ arrangement and inter-connections, functions, number of layers and hops, operating frequency bands and deployment environment, among others. It presents how RISs have evolved from passive to active and hybrid RIS models, from planar to multi-sector RISs, and from small-sized RISs to large and extra-large scale dimensions. It describes the evolution from regular, structured and centralized RISs to the irregular, unstructured/arbitrary and distributed RISs, respectively. This study also shows how RISs have evolved from single-layer, single-hop and single-function RISs to stacked, multi-hop and multi-functional RISs, respectively. Further, it extends to the newly-proposed stacked flexible intelligent metasurfaces that harness the benefits of both the stacked intelligent metasurfaces and the flexible intelligent metasurfaces. In terms of deployment, this work describes terrestrial RIS and non-terrestrial RISs for aerial, satellite and Martian environments, static and mobile RISs, radio frequency and optical RISs, as well as the transition from the traditional diagonal RIS to the beyond-diagonal RISs, referred to as the transition from RIS 1.0 to RIS 2.0. For each category of RIS in the taxonomy, this study describes the RIS model, overviews the relevant studies, and then compares and contrasts the different subclasses. It also clarifies the overlapping, albeit inconsistent, use of some RIS terminologies in the literature, and concludes with a discussion of the challenges and open issues that are subjects of ongoing research on the technology, including the future research directions. This contribution reveals that several early-stage descriptions of RIS as only reflective, passive and two-dimensional, for example, only hold for the specific subclasses and do not apply generally to all RIS types. The motivation for this treatise is to aggregate the wide array of architectures on the technology, harmonize the disjointed studies on the individual categories and subclasses, and to provide a beyond-state-of-the-art roadmap for the technology to extend the frontiers of RISs for next-generation networks.