A Fast Synthesis and Optimization Method for Filter Antenna Based on Neural Network

A novel optimization method for filter antenna design by use of neural network techniques with coupling matrix is proposed. This method can quickly and efficiently optimize the performance of filter antennas according to the customized demand criteria. Compared with traditional methods, this optimization method greatly simplifies the optimal design of filter antennas by avoiding the complex and time-consuming manual debugging of constructing model dimensions full-wave electromagnetic simulation. The model of T -patch filter antenna with the operating frequency of 2.45 GHz and the bandwidth of 0.3 GHz is used to demonstrate the effectiveness of the proposed method. Finally, the obtained optimization results are in high consistency with the expected index requirements.