The hype around reconfigurable intelligent surfaces (RIS) has escalated over the past five years. It began with communication theoretic studies based on “guesstimated” models, but gradually became more grounded through experimental validations in sub-6 GHz bands and the formation of an ETSI Industry Specification Group, which has laid the foundation for future standardization of RIS technology.

The outcomes of these efforts are mixed. On the one hand, it is clear that one can build and operate RIS roughly as envisioned. Several universities and companies have built functional RIS prototypes, and there are efficient channel estimation and beamforming algorithms for scenarios where the goal is to create a virtual line-of-sight (LOS) path around a block object (as illustrated in the figure).

On the other hand, it has become clear that practical hardware design is associated with signal losses that are often unaccounted for in theoretical studies, but limit the practical usefulness of RIS. Many actors in the telecom industry remain convinced that alternative solutions (e.g., network-controlled repeaters and small-cell base stations) are more attractive to deploy than RIS. This is aligned with my five-year-old assessment: “RIS is a hammer looking for a nail“. Since we have not yet found a practical problem that RIS can solve significantly better than other technologies, despite intensive research and pre-standardization work, I believe the RIS technology has passed the Peak of Inflated Expectations on the Gartner hype cycle curve.

I currently see the largest potential for RIS deployments in mmWave bands. This assessment is based on three observations:

1. The propagation losses through objects grow with the carrier frequency, making it more attractive to send signals around objects than through them at mmWave frequencies compared to sub-6 GHz bands. The coverage of mmWave base stations is typically limited to LOS, e.g., a room or a street segment. Hence, a RIS could extend that coverage to a neighboring room or crossing street by creating a virtual LOS path.
2. A RIS of a given physical size provides a larger gain at higher frequencies. The RIS collects incident signal energy proportionally to its size, so that part is frequency-independent. However, the directivity of the reflected beam (beamforming gain) increases with the carrier frequency, as more elements can be fitted into the aperture.
3. The deployment of small-cell base stations is less attractive at higher frequencies, as coverage is limited. One RIS can replace one base station in a mmWave band, whereas multiple RISs are needed to replace one small cell at lower frequencies, as the coverage area is larger from such a base station (e.g., it could cover multiple rooms or streets).

The deployment of 5G mmWave networks is currently very limited, but once we identify the right use cases for these bands, we should also consider deploying networks using a combination of base stations and RIS.

The following video explains the fundamentals of RIS and demonstrates how effectively it can improve coverage at mmWave frequencies. The experiments were made using the XRifle Dynamic RIS and Developer Kit from TMYTEK.