Expedited Design of Microstrip Antenna Subarrays Using Surrogate-Based Optimization

Computationally efficient simulation-driven design of microstrip antenna subarrays is presented. The proposed design approach aims at simultaneous adjustment of all relevant geometry parameters of the subarray, which allows us to take into account the effect of the feeding network on the subarray radiation pattern [in particular, the sidelobe level (SLL)]. In order to handle a large number of variables involved in the design process, a surrogate-based optimization procedure is proposed that exploits electromagnetic (EM) array simulations of variable fidelity and suitably designed correction techniques. The penalty function approach allows us to minimize SLL while maintaining array matching at a required level. The optimized designs are obtained at the low computational cost corresponding to a few high-fidelity subarray simulations. Experimental validation of the manufactured designs of microstrip subarrays operating at 10 GHz is also provided.