Design of broadband and dual-band microstrip antennas on a high-dielectric substrate using a genetic algorithm

The authors report the use of a genetic algorithm (GA) to design patch shapes for microstrip antennas on a high-dielectric substrate for broadband and dual-band applications. A full-wave electromagnetic solver is employed to predict the performance of microstrip antennas with arbitrary patch shapes. A two-dimensional chromosome is used to encode each patch shape into a binary map. A GA with a two-point crossover and geometrical filtering is implemented to achieve efficient optimisation. The resulting optimised patch structures for broadband and dual-band operations are described. For broadband application, the optimised patch antenna achieves a four-fold improvement in bandwidth when compared to a standard square microstrip antenna. For dual-band application, the optimised patch antenna shows a good impedance match at the design frequencies of 1 and 2 GHz. The GA-optimised designs are built on a FR-4 substrate and measurement results show good agreement with the numerical prediction. Physical interpretation for the optimised shapes is discussed.