From intent to an antenna model

LADS connects literature-based antenna reconstruction with parameter optimization. LEAM is its open-source modeling component. The paper demonstrates a CST workflow in which descriptions and figures become parameterized solids, executable macros, and an SB-SADEA optimization setup.

From literature to a simulation model

01Identify solidsInterpret descriptions and figures
02Define solidsDimensions
03Execute macrosBuild and check the CST model
04Optimize parametersConfigure and run SB-SADEA

↻ Engineers inspect intermediate artifacts and review design changes. Method summary redrawn from the paper.

The system includes 11 LLM tools. Its UWB example reconstructs a cross-slotted monopole, then changes the geometry to an H-slot and the substrate to Rogers RT 5880, with engineers selecting and reviewing the modifications.

Geometry and simulated response

Parameterized H-slotted monopole geometry — The generated H-slot design has 12 design variables. Paper Fig. 5 ↗

Simulated reflection coefficient and realized gain of the H-slotted antenna — Reflection coefficient and realized gain of the optimized design; these are simulated responses. Paper Fig. 6 ↗

3.1–10.6 GHztarget UWB band

0.99 dBigain variation, versus 1.71 dBi in the reference

721EM evaluations; about 12 hours

A feasible design was found after 256 evaluations. The optimization then continued to reduce gain variation while maintaining the gain level.

Paper and implementation

The published demonstration uses CST. The current LEAM toolkit also offers HFSS / PyAEDT modeling workflows; these extensions should be distinguished from the scope validated in the paper. Documentation.