Capacity for Electromagnetic Information Theory

Traditional channel capacity based on one-dimensional time domain mismatches the four-dimensional electromagnetic fields, thus it cannot fully exploit the information in the spatial dimensions. Therefore, electromagnetic information theory based on the four-dimensional electromagnetic fields becomes necessary to reveal the fundamental theoretical capacity bound of the communication systems. Existing works on electromagnetic information theory focused on deterministic signals and degrees of freedom, which were unable to derive the capacity due to the lack of entropy definition. In this paper, we first model the communication between two continuous regions by random field. Then, we analyze a special case with parallel linear source and destination to derive the capacity bound. Specifically, for parallel infinite-length source and destination, we analyze the mutual information by spatial spectral density and derive the best current distribution on the source to achieve the maximum mutual information, i.e., the capacity. Then, we analyze the scenario with infinite-length source and finite-length destination. We use Mercer expansion to derive the mutual information between the source and the destination. Finally, for a practical model with finite-length source and destination, we analyze its Mercer expansion and reveal its connection with the infinite-length case.