Deep Learning Assisted Automatic Methodology for Implanted MIMO Antenna Designs on Large Ground Plane

This paper provides a novel methodology for designing implanted multiple-input and multiple-output (MIMO) antennas in the automatic fashion. The proposed optimization consists of two sequential phases for firstly configuring the geometry of an implanted MIMO antenna and then sizing the design parameters through the hierarchy top-down optimization (TDO) and regression deep neural network (DNN), respectively. It tackles the difficulty in constructing the structure of antennas and also provides optimal values for the determined variables, sufficiently. This methodology results in valid electromagnetic (EM)-verified post-layout generation that is ready-to-fabricate. The effectiveness of the proposed optimization-oriented method is verified by designing and optimizing the implanted MIMO antenna in the frequency band of 4.34–4.61 GHz and 5.86–6.64 GHz suitable for medical applications at the emerging wireless band. For our design, we employ the actual biological tissues as bone, liquid (%1 sodium chloride, @ sugar in distilled water), and plexiglass surroundings with a bio-compatible substrate, as aluminium oxide on a large ground plane, that is suitable to be used in a particular biomedical applications involving smart implants.