A Study On The Evolutionary Optimization Of Planar Antennas

Antennas with planar structure are extensively utilized in modern wireless communication systems because of their attractive features including low profile, small size and light weight. With the developing of electromagnetics computer-aided design tools, optimizing the design of antennas using evolution algorithms, such as genetic algorithms, is a focus.A plug-in optimizer based on Python programming and HFSS simulation software for antenna optimization is presented in this thesis. And team progress algorithm (TPA) is used as optimization algorithm. The objective function is established to optimize the impedance bandwidth of some specific antennas.To avoid use time-consuming EM simulator HFSS in evolution algorithm, the error back propagation neural network (BPNN) is employed to establish the nonlinear relationships between the antenna structure parameters and the performance parameters. BPNN can be trained by using samples generated by HFSS. Therefore, computational efficiency can be obviously improved by the trained BPNN and TPA.Two ultra-wideband printed monopole antennas are optimized through plug-in optimizer, simulation software and BPNN in this thesis. One has lower return loss in its working bandwidth after optimized, the other achieves the different requirements through different objective functions.The feasibility of the optimization method is proved by there antenna optimization examples. The plug-in optimizer and BPNN, which is presented in this thesis, provide a technical way to optimize antennas use commercial software HFSS and users optimization algorithms.