A novel metamaterial power divider optimized by artificial neural network

An approach for enhancing the operating bandwidth of classic dual-band power divider (PD) is proposed by using metamaterial (MTM) units. To overcome limitation EM simulation and improve design efficiency, we propose an artificial neural network (ANN) that enables inverse prediction MTM-PD geometry from its desired physical response. In ANN approach, convolutional long short-term memory are combined to learn relationship between geometric corresponding responses then accurately predict parameters MTM-PD. The predicted includes low frequency (LF) band 1.90–2.43 GHz high (HF) 4.61–5.25 GHz. Compared PD, LF has been enhanced five times. measured results confirm both HF with a 0.5 GHz, verifying reliability our approach. It demonstrates potential designing microwave devices solving electromagnetic problems.