An Automated Machine Learning Approach to Inkjet Printed Component Analysis - A Step Toward Smart Additive Manufacturing

Aug 2, 2026

Upload by Ezra

Machine learning is used to automate microwave characterization of inkjet printed components on flexible substrates, enabling simultaneous extraction of ink conductivity and dielectric properties. The approach trains models using the mutual dependence between inkjet printed coplanar waveguide material parameters and EM-simulated propagation constants, then combines the learned models with measured propagation constants to recover prototype parameters. Four heuristic machine learning models are compared to validate performance, with eXtreme Gradient Boosted Trees (XGB) and Light Gradient Boosting (LGB) providing the best characterization results under the studied conditions.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages6

Reading Time15 min

AccessFree to download

Introduction

Data Extraction

Fabrication and Measurement

Machine Learning Models

Material Characterization and Model Comparison

Conclusion

An Automated Machine Learning Approach to Inkjet Printed Component Analysis: A Step Toward Smart Additive Manufacturing Abhishek Sahu 1 , Peter H. Aaen2 and Praveen Damacharla3 1 Qorvo Us Inc., Greensboro, NC, USA 2 Colorado School of Mines, Golden, CO, USA 3 KineticAI Inc., The Woodlands, TX, USA [1](1 abhishek.sahu@ieee.org)[ abhishek.sahu@ieee.org](1 abhishek.sahu@ieee.org), [2](2 paaen@mines.edu)[ paaen@mines.edu](2 paaen@mines.edu), [3](3 praveen@kineticai.com)[ praveen@kineticai.com](3 praveen@kineticai.com), arXiv :2404 .04623v 1 [ cs .LG] 6 Apr 2024 Abstract—In this paper, we present a machine learning based architecture for microwave characterization of inkjet printed components on flexible substrates. Our proposed architecture uses several machine learning algorithms and automatically selects the best algorithm to extract the material parameters (ink conductivity and dielectric properties) from on-wafer measurements. Initially, the mutual dependence between material parameters of the inkjet printed coplanar waveguides (CPWs) and EM-simulated propagation constants is utilized to train the machine learning models. Next, these machine learning models along with measured propagation constants are used to extract the ink conductivity and dielectric properties of the test prototypes. To demonstrate the applicability of our proposed approach, we compare and contrast four heuristic based machine learning models. It is shown that eXtreme Gradient Boosted Trees Regressor (XGB) and Light Gradient Boosting (LGB) algorithms perform best for the characterization problem under study. Index Terms—AutoML, Inkjet printing, LightGBM, Printed electronics, ResNet, XGBoost I. INTRODUCTION Printed electronics has gathered wide attention in the emerging markets such as internet of things as it offers biodegradable and cost-effective solutions. Recently, many traditional methods such as gravure printing, flexography, screen printing, and inkjet printing have been introduced in manufacturing of microwave circuits [1] . Among them, inkjet printing is most popular in the scientific and industrial community due to its attractive features including low manufacturing cost, large-area processability and lower carbon footprints. RF/microwave community has reported numerous applications of inkjet printing including antennas, wireless power transfer topologies, sensors, and microwave components [2]–[5] for low-loss and high-speed communication systems. Significant effort has been devoted from the research community to extract the material parameters of inkjet printed components from their measurements [6] - [10] . Conventional approaches need high precision profile measurements to extract the ink conductivity and a conformal mapping method (CMM) based on elliptical integrals to evaluate the dielectric constants. However, as these analytical methods are manually calculated to extract the parameters independently, they require rigorous human effort. Recently, computer-aided Fig. 1. Measurement setup for the inkjet printed CPWs. GSG probes are connected to measured the small signal S-parameters of the printed CPWs. An LCR meter is connected in series to measure the dc resistance of the THRU line. design (CAD) based-techniques utilizing commercial EM simulation packages and measurements have been proposed for automation and concurrent parameter extraction [10]–[13] . The problem with this approach is it requires a large number of iterative simulations. Further, as the number of variables increase, the optimization space becomes very complex and sometimes may lead to convergence problems. Therefore, there is a need for an alternative characterization approach that offers a more straightforward automation and is able to extract multiple parameters simultaneously without increasing process complexity. Recently machine learning based approaches have gained wide attention as they make the characterization problem computationally int

FAQs

What problem does the proposed AutoML approach address for inkjet printed components?

It automates microwave characterization by extracting ink conductivity and dielectric properties from on-wafer measurements without requiring extensive manual algorithm selection or complex iterative simulation workflows.

How are the machine learning models trained in the proposed architecture?

Models are trained using the mutual dependence between inkjet printed CPW material parameters and EM-simulated propagation constants, establishing a relationship between measurable RF/microwave behavior and target material properties.

Which machine learning algorithms perform best in the characterization study?

The eXtreme Gradient Boosted Trees (XGB) and Light Gradient Boosting (LGB/LightGBM) regressors achieve the strongest performance for the studied characterization problem.

Aug 2, 2026

Upload by Ezra