Knowledge based EM-ANN models for the design of wide bandwidth CPW patch/slot antennas

Abstract

Knowledge based electromagnetically-trained artificial neural network (EM-ANN) models have been developed for the design of wide bandwidth CPW patch/slot antennas fed at the radiating edge. The knowledge based EM-ANN modeling technique is aimed at improving the accuracy of an existing model or improving/extending the existing model in some way. Existing models are used as prior knowledge in order to reduce the complexity of the new model to be developed. An existing first order model for the CPW patch/slot antenna has previously been developed using a segmentation approach in order to reduce computation time. This paper utilizes knowledge based ANN modeling techniques to enhance the existing CPW patch/slot antenna model by including second order effects, such as coupling between segmented pieces, and to alter the antenna feed structure for improved impedance matching. A wide-bandwidth antenna design has been carried our near 5.4 GHz. The developed modeling approach is general and can be used for modeling other antenna structures as well.