Wideband Microstrip Antenna Design Based On NSGA-Ⅱ And Neural Network

Since ultra-wideband wireless technology is developing rapidly and being widelyused, tough requirements are demanded in the design of antenna when applied in theUWB communication system. Therefore, researchers worldwide are focusing on howto design a compact microstrip antenna with high performance. Genetic algorithm,which is based on the natural selection mechanics, belongs to the class of stochasticalgorithms. It is generally applied in antenna optimization design. However, thealgorithm has its own disadvantages: first of all, it is inevitably difficult to programmethe complex structures of some antennas and always time consuming if we turn toelectromagnetic simulation software to calculate the value of the fitness function.Moreover, tradition algorithms normally weighted all values of the objective functionsto obtain the value of the fitness function, and then continue the optimization processwith reference to this parameter. Evidently, it is impossible to comprehensively assessall aspects of the antenna’s performance.In this paper, an automatic optimization method aimed at improving theperformance of UWB antenna combined neural network, NSGA-II algorithm withHFSS is proposed. We establish the relationship between the antenna structureparameters with its performance results through the neural network; use the non-dominated sorting genetic algorithm to value the desired indicators such as return loss,radiation pattern and gain; Finally we obtain the best structure of the targeted antenna.In this paper, a detail description is presented about the main modules, programfiles and whole procedure included in the optimization proposal. We reduce the returnloss of an UWB antenna to lower than-10dB over4~6GHz band frequency by theproposed optimization method. When it comes to changing the shape of radiated patchto improve its performance, we offer a novo way to cut slots in the patch, and effectivelybroaden the bandwidth of a microstrip antenna. In the last chapter, we improved theperformance of a U-shape monopole antenna by changing its shape. The antenna wasmanufactured, its measured performance results showed perfect match with thesimulation results. In addition, we also designed a CPW UWB antenna, a UWB antennawith symmetrical circle structure, and an antenna with dual band-notchedcharacteristics, all the above antennas have been resolved by HFSS, the simulationresults showed that they all meet the design purpose of broad bandwidth and compactsize.