Benchmarking Framework for Reconfigurable Intelligent Surfaces

Reconfigurable Intelligent Surfaces (RIS), which comprise of large number of unit cells (meta-atoms), reflect the incoming wave according to desired radiation patterns. As the number of unit cells increase, complex radiation patterns can be generated. Similarly, a unit cell allowing more phase control and larger phase difference helps in the generation of complex radiation patterns at the surface level. However, the control circuit complexity and energy requirements also increase. In many practical applications, only a handful of radiation patterns may suffice. Additionally, the overall problem of finding the appropriate unit cell control state to generate any desired radiation patterns has combinatorial complexity. In this paper, we propose a benchmarking framework and suitable performance metrics that enable us to determine and compare the capabilities of RISs made from different unit cells. The proposed framework is numerically tested on three 40x40 RISs made from optimized and unoptimized unit cells of various resolutions reported in the literature. While the performance of RIS made from 1-bit unoptimized unit cell was overall poor, it was able to successfully generate some simple but useful benchmarking patterns. Thus, our framework provides a mechanism to identify the best candidate unit cell for a given set of requirements. The proposed framework has a potential to revolutionize future research and development on unit cell and RIS design.