Related Experiment Video
Updated: Sep 11, 2025
09:12
Colloidal Synthesis of Nanopatch Antennas for Applications in Plasmonics and Nanophotonics
Published on: May 28, 2016
11.3K
Low-cost fabrication and comparative evaluation of machine learning algorithms for flexible PDMS-based hexagonal
Srivatsan Sarvesan1, Mettu Goutham Reddy2, S S Karthikeyan3
1Department of Electronics and Communication Engineering, NIT Tiruchirappalli, Tiruchirappalli, 620015, India.
Scientific Reports
|August 18, 2025
Summary
Researchers developed a flexible hexagonal antenna on Polydimethylsiloxane (PDMS) for 5G and wearables. Machine learning optimized design, achieving high accuracy for improved impedance matching and consistent performance in a functional prototype.
Area of Science:
- Electrical Engineering
- Materials Science
- Machine Learning
Background:
- Flexible antennas are crucial for emerging 5G and wearable technologies.
- Mechanical bending can significantly degrade antenna performance.
- Efficient design and fabrication methods are needed for practical realization.
Purpose of the Study:
- To design and develop a flexible hexagonal microstrip patch antenna on a Polydimethylsiloxane (PDMS) substrate.
- To investigate the use of machine learning for optimizing antenna design parameters.
- To demonstrate a low-cost fabrication method and validate performance through prototyping and circuit modeling.
Main Methods:
- Utilized a hexagonal microstrip patch antenna design operating in a higher-order resonant mode.
- Employed four supervised machine learning algorithms (Random Forest, XGBoost, CatBoost, LightGBM) trained on electromagnetic simulation data.
- Developed a scalable fabrication process for conductive traces on PDMS and created a functional prototype.
Main Results:
- The Random Forest model achieved the highest predictive accuracy (R² = 0.99) for optimizing geometrical parameters.
- The fabricated flexible antenna prototype demonstrated consistent performance with a measured gain of 3.2 dB.
- An electrical equivalent lumped-element circuit model validated the electromagnetic simulation results.
Conclusions:
- Machine learning significantly accelerates the design and optimization of flexible antennas.
- The developed fabrication technique is cost-effective and scalable for PDMS-based antennas.
- This integrated approach facilitates the practical development of high-performance flexible RF antennas for 5G and wearable applications.
