New Approaches for a Reconfigurable Microstrip Patch Antenna Using Inverse Artificial Neural Networks

Inverse artificial neural networks (ANNs) are used in this study to design and improve a reconfigurable 5-fingers-shaped microstrip patch antenna. Utilizing three precise prior knowledge inverse ANNs and a sizable quantity of training data, new solutions are created by including frequency information into the design of the ANNs. The recommended antenna resonates at frequencies that range from 2 to 7 GHz and may configure in four modes, each of which is controllable by two PIN diode switches featuring ON/OFF states. Utilizing prior knowledge mitigates the complexity of the input/output collaboration. With a multilayer perceptron (MLP) as the first phase of the training process, three separate techniques of knowledge incorporation are illustrated, and their outcomes are compared to those of an EM simulation.