MIMO: Does It Make Sense From an Electromagnetic Perspective and Illustrated Using Computational Electromagnetics?

The objective of this article is to illustrate the concepts of multiple-input-multiple-output (MIMO) antenna systems from an electromagnetic perspective and the principles illustrated using computational techniques. The various conceptual evolutionary steps of MIMO have been described. This methodology went through various distinct, disjoint philosophies to satisfy its critics as the presented techniques had various shortcomings from a physics perspective. This is because the principle of superposition of power is not always valid in electrical engineering and hence not applicable to analysis of multiantenna systems, where multistreams are affected by mutual coupling. The prime objective of this article is to specify that the input power needs to be the same when comparing performance between systems, as it is not clear how one excites an antenna using a priori fixed power, as antennas are excited with voltage sources. Also, these misguided researchers consider an antenna to radiate power as they are often treated as point sources. In reality, an antenna radiates an electric/magnetic field. Using computational electromagnetics, it is illustrated that a 1 × 1 single-input-single-output (SISO) system may perform better than a 2 × 2 MIMO system under specific circumstances. However, as the concept of channel capacity is developed on purely mathematical grounds based on statistics and not on physics of electromagnetics, it is difficult to interpret its results. Also, Shannon developed his formula when coding at the RF stage and not in the baseband, which is often missed. The channel capacity, in addition, is defined with respect to background thermal noise, whereas no receiver can accept a signal weaker than 100 μV/m in the absence of interference, which is far above the background noise and hence unrealistic. Finally, a zero-forcing scheme based on electromagnetic principle of reciprocity is presented. Numerical examples using computational electromagnetics are presented to illustrate this scenario and a comparison in performance with conventional systems is made illustrating that this system can perform well.