A New COVID-19 Detection Method Based on CSK/QAM Visible Light Communication and Machine Learning.

Area of Science:

• Biomedical Engineering
• Optical Communications
• Machine Learning

Background:

• Accurate and rapid detection of COVID-19 remains critical.
• Visible Light Communication (VLC) offers a novel channel for data transmission.
• Machine Learning (ML) provides powerful tools for pattern recognition and classification.

Purpose of the Study:

• To propose and evaluate a novel method for COVID-19 detection using VLC and ML.
• To model COVID-19 DNA gene transfer within a CSK/QAM-based VLC system.
• To identify the optimal ML model for classifying COVID-19 samples.

Main Methods:

• Development of mathematical models for COVID-19 DNA gene transfer in square constellations.
• Application of ML algorithms (including XGBoots) for classifying electrophoresis samples.
• Performance analysis based on Bit Error Rate (BER) and constellation complexity.

Main Results:

• The XGBoots model achieved the highest accuracy (96.03%) and recall (99%) for positive COVID-19 samples.
• Complexity studies indicated optimal performance for the N=2^2i×2^2i, (i=3) square constellation.
• Performance analysis showed significant gains in signal quality for various constellation sizes at BER = 10^-3.

Conclusions:

• The proposed VLC and ML integrated system demonstrates high efficacy for COVID-19 detection.
• XGBoots is identified as the superior ML model for this specific application.
• The study highlights the potential of optical communication systems in medical diagnostics.