Reconfigurable Electromagnetic Environments: A Signal Processing Approach

The emergence of smart radio environments (SREs) and holographic radio paradigms, which rely on reconfigurable intelligent surfaces (RISs), has highlighted the importance of developing physically consistent models and design tools for communication systems [1, 2]. As we move towards a future characterized by reconfigurable electromagnetic (EM) environments, the primary objective will be to optimize the system response in real time by adjusting the parameters of the electromagnetic objects (EMOs) composing the system, such as the reflection properties of metasurface-based RISs. Achieving this goal requires a holistic system view that emphasizes the need for tractable and physically consistent design tools that integrate both signal processing and EM theory [3]. In this perspective, starting from rigorous EM arguments, we have developed a comprehensive framework for the characterization and design of programmable EM environments that is valid in both the far-field and near-field regimes (radiative and reactive). This framework is designed to provide a system-theoretic and physically-consistent interpretation of reconfigurable EM environments. Our approach reveals that any system involving linear EMOs can be viewed as a space-variant feedback filter (see Fig. 1). Specifically, we demonstrate that boundary conditions applied to a generic EMO can be represented as a feedback system. Analyzing and designing space-variant feedback systems can be challenging, so we developed an approach that incorporates linear algebra operations over a graph to design and characterize EM systems with reconfigurable EMOs. We also establish a relationship between the EM transfer function and the linear algebra representation of the system, with a focus on characterizing intelligent surfaces.