A Multi-Substrate Dataset for Modified Slotted Microstrip Patch Antenna as Electrolyte Sen

5,625 finite element simulation records for a microstrip patch antenna sensor designed for non-invasive sweat electrolyte monitoring at 2.4 GHz. The dataset, created by Rakib, Sakhawat Hossen and hosted on Harvard Dataverse, was generated using COMSOL Multiphysics 6.3 and includes simulations across three substrate materials and five NaCl concentration levels. It was last updated on June 1, 2026.

Use Cases

• Train surrogate machine learning models for antenna design optimization based on geometric and electrolyte input parameters.
• Analyze the relationship between substrate material and sensor performance metrics like reflection coefficient and resonant frequency.
• Benchmark regression algorithms (e.g., XGBoost, SVR, GPR) on synthetic electromagnetic simulation data.
• Investigate the dielectric sensitivity of antenna sensors to physiologically relevant NaCl concentrations.

Strengths

• Contains 5,625 simulation records generated from a systematic grid of input parameters.
• Covers three distinct substrate materials (Rogers RO4003C, FR4, Cellulose Paper) and five NaCl concentration levels.
• Includes a structured 80/20 train-test split and model performance comparisons, with reported XGBoost RMSE for S11 (0.81-1.14 dB) and concentration (1.31-1.82 mmol/L).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but the specific data format and sample structure require inspection upon download.
• Data is synthetic from simulations; real-world performance may differ.

Provenance

Source

Harvard Dataverse

Collection Method

Generated via finite element simulations using COMSOL Multiphysics 6.3 (RF Module).

Freshness

Last updated 2026-06-01 20:28:40; freshness should be verified.