Traditional massive multiple-input multiple-output (MIMO) information theory adopt non-physically consistent assumptions, including white-noised, scalar-quantity, far-field, discretized, and monochromatic EM fields, which mismatch the nature of the underlying electromagnetic (EM) fields supporting the physical layer of wireless communication systems. To incorporate EM laws into designing procedures of the physical layer, we first propose the novel concept of EM physical layer, whose backbone theory is called EM information theory (EIT). In this article, we systematically investigate the basic ideas and main results of EIT. First, we review the fundamental analytical tools of classical information theory and EM theory. Then, we introduce the modeling and analysis methodologies of EIT, including continuous field modeling, degrees of freedom, and mutual information analyses. Several EIT-inspired applications are discussed to illustrate how EIT guides the design of practical wireless systems. Finally, we point out the open problems of EIT, where further research efforts are required for EIT to construct a unified interdisciplinary theory.

翻译：传统的大规模多输入多输出（MIMO）信息论采用非物理一致的假设，包括白噪声、标量量、远场、离散化和单色电磁场，这些假设与支撑无线通信系统物理层的底层电磁场的本质不匹配。为将电磁定律融入物理层设计流程，我们首先提出电磁物理层的新概念，其核心理论称为电磁信息论（EIT）。本文系统性地研究了EIT的基本思想与主要成果。首先，回顾经典信息论与电磁理论的基本分析工具；其次，介绍EIT的建模与分析方法，包括连续场建模、自由度与互信息分析；接着，讨论若干EIT启发的应用案例，阐述EIT如何指导实际无线系统的设计；最后，指出EIT的开放问题，需进一步研究以构建统一的跨学科理论。