Adaptive Neuro-Fuzzy Inference System approach in bandwidth and mutual coupling analyses of a novel UWB MIMO antenna with notch bands applicable for massive MIMOs

A novel UWB MIMO antenna with band-notched characteristic operating in the frequency range between 3.1 GHz and 10.6 GHz is presented. The designed MIMO antenna has two ports and the size of 32×14 mm2 which is fabricated on an FR-4 printed-circuit-board. The feeding system of the proposed antenna is a microstrip line. The two antennas are positioned face to face but laid reversely upon the substrate in order to have a good isolation, less cross polarization, high gain and good envelope correlation coefficient. S-parameters describe the input-output relationship between ports (or terminals) in a MIMO antenna. The effects of the strip feed line width on the variations in reflection coefficient (S11) and consequently the impedance bandwidth as well as the mutual coupling between the two antennas for various separations are analyzed by Adaptive Neuro-Fuzzy Inference System (ANFIS) approach. This system is trained for different membership functions based on the number of epochs, root-mean-square error and mean absolute percentage error. Then, a comparative study is made between them. It is shown that when fuzzy system is trained by gaussian combination membership function (Gauss2mf), better results in terms of training, testing error and correlation coefficient can be attained.