Modeling of Mobile Antenna optimization based on Artificial Neural Network

Artificial Neural Network (ANN) has been realized as a robust technique that successfully applied in many Electromagnetic problems. Based on the learning process from prior knowledge, ANN can be considered as a non-linear mapping of the complex relationship of the inputs and the outputs. This reliable replacement model (referred to surrogate model) allows reducing the computational cost of high-EM simulation in antenna designs. In this article, we propose a novel approach where ANN is trained with the data which is collected from mobile antenna simulation to retrieve a surrogate model. This work contributes a new model for researchers to save time and efforts in mobile antenna design because results will be predicted quickly and accurately by the model instead of by the computational expensive cost of simulation. To be more specific, engineers can tune the antenna in stand-alone to get new data and input them into the predicted model to observe |S11| of the antenna in mobile.