AI and IoT-Enabled Multi-Sensor CNN Model for Non-Invasive Blood Viscosity Estimation

View at publisher Summarise with VIRAT
Authors: R Saravanakumar, Mangai S, Mithanya M, Mowniga R, Pugalnathan S, Vinodha R
Cited by: 0
References: 19

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Continuous monitoring of Blood Viscosity (BV) is important for the early diagnosis of Cardiovascular Diseases. Traditional methods of BV measurement are mainly invasive and based on complex, costly, non-clinical equipment. However, the non-invasive methods available provide only the average viscosity value with limited accuracy. To overcome these limitations, this project proposes a novel AI-based noninvasive estimation of BV using Photoplethysmography, Bio-impedance, and Temperature sensors. The PPG provides waveform features associated with blood flow and perfusion, Bio-impedance reflects electrical properties related to cellular concentration, while the Temperature provides a physiological compensation factor. This paper proposes a CNN model to estimate BV using filtered PPG signals, which are processed through a Butterworth-type FIR filter. The data is taken from the MIMIC II database in the Physionet Repository. The proposed model performs a high prediction accuracy of 99.87% for 0.0025 loss. Additionally, an IoT-based patient monitoring system using ESP8266 that continuously acquires health parameters and broadcasts them over a dedicated Wi-Fi IP address is implemented. Subsequently, a prototype has been developed to validate system performance. This integrated system can be used to realize real-time, non-invasive blood viscosity monitoring, timely clinical decisions, and regular health assessment in patients.

Fields of Study & Subjects

Non-Invasive Vital Sign Monitoring Hemodynamic Monitoring and Therapy Artificial Intelligence in Healthcare

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
457–461
Publication date
21 January 2026
Language
EN