A meandered dual-band power divider with improved performance for wireless applications using a neural network inverted model

In this paper, a dual-band Wilkinson power divider is presented using open-stubs and meandered lines. The first operating frequency of this divider is 2 GHz, and the second operating frequency is 4 GHz. An artificial neural network model is proposed to improve the divider performance and help to find the optimized size of the applied stubs. The proposed inverted neural network model can predict dimensions of the power divider based on the desired performance. The designed planar divider features a simple structure with an easy fabrication process. This divider is designed, simulated, and fabricated. The presented circuit is fabricated on an RT-5880 substrate with a dielectric constant of 2.2 and a thickness of 0.508 mm. The overall dimensions of this circuit are 25.4 mm × 27.1 mm. Compared to the previous dual-band power dividers, the proposed divider shows improved performance at both operating frequencies with low insertion loss, high output port isolation, and desirable return loss. To verify the correct performance of the presented approach, the measurement results are provided. The measured insertion loss, input return loss, output return loss, and output port isolation parameters are obtained as 0.6 dB, 26.5 dB, 26 dB, and 25 dB at 2 GHz, while they are 0.6 dB, 19 dB, 24.5 dB, and 16 dB at 4 GHz operating band.