Design Optimization of Planar Structures Using Self-Adjoint Sensitivity Analysis

A new practical approach to the $S$-parameter sensitivity analysis of planar structures is proposed. The response sensitivities are obtained with a self-adjoint approach with discrete shape perturbations on non-uniform method-of-moment grid. The technique is illustrated through the $S$-parameter sensitivity analysis of a planar printed antenna and a microwave double-stub tuner. The computed sensitivities are validated by comparing with the central finite-difference estimates at the response level. Gradient-based optimization is performed using the sensitivity information. The performance of the optimized design is verified by simulation and by measurement.