Nonlinear Mutual Coupling Compensation Operator Design Using a Novel Electromagnetic Machine Learning Paradigm

We propose a technology that utilizes a unified electromagnetic machine learning (EM-ML) technique to mitigate the effect of mutual coupling in receiving antenna array configurations. The recently developed antenna current Green's function (ACGF) formalism is deployed to explicate the electromagnetic behavior of antennas in the form of an accurate digital signal processing (DSP) model, including mutual coupling interactions between radiators. A deep learning framework is devised and combined with the ACGF-based DSP model to design a novel nonlinear mutual coupling compensation operator providing higher decoupling capabilities in comparison to previously reported linear methods. A direction-of-arrival estimation application is presented to validate the proposed EM-ML system.