A novel microwave sensor for multi-component liquid analysis using machine learning-based edge computing

G. Challa Ram

doi:10.1016/j.measurement.2025.118531

A novel microwave sensor for multi-component liquid analysis using machine learning-based edge computing

2025-08-04

0

PRE

AI

DOI：10.1016/j.measurement.2025.118531

Original

Original request for help

Save

Abstract

En 中文

• Proposes a novel non-invasive microwave sensor using a spoof surface-based whispering-gallery mode resonator for multi-liquid component analysis. • Demonstrates accurate volumetric concentration detection in mixtures of five mutually soluble liquids using spectral response analysis. • Employs a multivariable regression-based machine learning model with resonance features for precise concentration prediction. • Integrates principal component analysis (PCA) to enhance feature selection, achieving a low RMSE of 0.025. • Features a Raspberry Pi 4-based edge computing system enabling real-time, automated analysis in a portable lab-friendly setup.

Keywords:

microwave sensor

whispering-gallery mode

multivariable regression

principal component analysis

edge computing

A novel microwave sensor for multi-component liquid analysis using machine learning-based edge computing

Abstract

Journal

Organization

Cited Papers