Using Machine Learning for Mutual Coupling Compensation in Strongly Coupled Arrays

Recent advances in radio frequency (RF) and baseband integrated circuits technology has forced a high demand on the size minimization of array-based microwave systems. Conversely, minimization of array aperture consequently causes a strong mutual coupling in the array systems. Presently, estimation and compensation of mutual coupling (with nonlinear profile) is still a technical problem that requires research attention. For example, realization of microwave orientations with unknown mutual coupling remains a challenge. In order to address this issue, a unified electromagnetic machine learning (EMML) technique is presented in this chapter. By including the impact of mutual coupling among the components, the digital signal processing (DSP) module uses the antenna current's Green function (ACGF) to characterize the electromagnetic response of the elements. In order to create a nonlinear mutual coupling mitigation scheme that provides superior decoupling capabilities over earlier linear approaches, an EMML scheme is then designed and combined with the DSP module based on ACGF. A comparison with conventional methods was carried out after the suggested EMML was validated using a direction-of-arrival estimation demonstration based on multiple signal classification (MUSIC). The results demonstrate the effectiveness and validity of the suggested EMML approach. This method can be applied to various scenarios such as coupled array-based beamforming, steering, and nulling, and has applications in antenna design, metamaterials, wireless communications, and wave propagation in complex environments.