Reconstruction Design of Semi-Automated Antenna Topology Based on Joint Simulation Drive

This study addresses the computational challenges posed by diversified hexagram-inspired unit cell combinations in antenna topology optimization and the inefficiency of traditional ANSYS-based simulations by proposing a semi-automatic inverse design method using MATLAB-ANSYS co-simulation. Binary matrices represent “disconnected” and “connected” unit states, generating 64 configurations from eight fundamental arrangements. A D4 symmetry constraint ($x-y$ axis mirror symmetry) ensures maximum gain along the main axis in a $4 \times 4$ metamaterial structure. Three MATLAB-HFSS co-driven programs are developed: (1) a parametric HFSS modeling module automates script generation, improving efficiency by $60 \%$; (2) a distributed scheduler enables parallel multi-core simulation, reducing analysis time from 32 to 10.5 hours; (3) a probabilistic histogram-based inverse prediction rapidly identifies optimal configurations via gain-parameter response surfaces and iterative conditional probabilities. A substrate-gain SPSS correlation database further identifies the optimal substrate layout. Experiments show the method maintains directivity above 11 dB, reduces labor costs near zero, shortens design cycles to $\mathbf{1 0. 2}$ hours, and supports scaling to larger arrays. This framework efficiently facilitates inverse topology optimization for large-scale antenna design by pinpointing substrate configurations with the strongest gain correlation.