Perancangan Wearable Healthvest Pemantau Tanda Vital Multi-Parameter Berbasis Sinyal Elektrokardiogram Dan Fotoplestimogram Menggunakan Advanced Signal Processing

Pranata, Aldo (2025) Perancangan Wearable Healthvest Pemantau Tanda Vital Multi-Parameter Berbasis Sinyal Elektrokardiogram Dan Fotoplestimogram Menggunakan Advanced Signal Processing. Other thesis, Institut Teknologi Sepuluh Nopember.

This is the latest version of this item.

[thumbnail of 5023211019-Undergraduate_Thesis.pdf] Text
5023211019-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only
Download (29MB)

Abstract

Penelitian ini mengembangkan perangkat wearable berupa healthvest yang dirancang untuk pemantauan kesehatan tubuh secara berkelanjutan menggunakan kombinasi sensor electrocardiogram dan photoplethysmogram. Pemantauan tanda vital tubuh sangat penting untuk deteksi dini gangguan kesehatan, khususnya penyakit kardiovaskular. Metode konvensional seperti pemeriksaan di fasilitas kesehatan memiliki keterbatasan dalam hal frekuensi pemantauan, aksesibilitas, dan efisiensi biaya. Sementara itu, teknologi wearable komersial seperti smartwatch menawarkan kemudahan, namun umumnya terbatas dalam jumlah parameter yang dapat dipantau akibat keterbatasan ukuran dan desain perangkat. Healthvest yang dikembangkan dirancang untuk mengatasi keterbatasan tersebut dengan solusi multi-parameter yang nyaman, stabil, dan mudah diakses. Sistem terdiri dari elektroda strap-electrocardiograph yang dijahit pada bagian dalam vest untuk menjaga kestabilan sinyal serta unit sensor photoplethysmograph pada belt unit. Dari sinyal electrocardiogram dilakukan analisis heart rate variability menggunakan metode discrete wavelet transform untuk ekstraksi R-R interval, diikuti analisis domain waktu dan estimasi laju pernapasan melalui komponen frekuensi tinggi. Sinyal photoplethysmogram digunakan untuk estimasi kadar oksigen dalam darah dengan pendekatan machine learning berbasis random forest serta prediksi kadar glukosa darah dan tekanan darah menggunakan model deep learning berbasis convolutional neural network dan temporal convolutional network yang dilengkapi convolutional block attention module (CBAM) . Seluruh data diproses secara lokal dan dikirim ke server secara nirkabel berlatensi rendah (<1 ms), lalu ditampilkan melalui antarmuka pengguna yang ramah. Hasil pengujian menunjukkan perangkat ini mampu menghasilkan sinyal stabil dan estimasi tanda vital yang akurat, dengan mean absolute error (MAE) sebesar 4.59 bpm untuk denyut jantung, 1.15 brpm untuk laju pernapasan, 1.13% untuk SpO2, 14.275 mg/dL untuk kadar glukosa darah, serta 6.22 mmHg dan 4.46 mmHg untuk tekanan darah sistolik dan diastolik. Perangkat ini berpotensi menjadi solusi pemantauan kesehatan harian yang lebih komprehensif dan nyaman.
=========================================================================================================================================
This research develops a wearable health monitoring device in the form of a healthvest, designed for continuous monitoring of vital signs using a combination of electrocardiograph and photoplethysmograph sensors. Monitoring vital signs is essential for the early detection of health problems, particularly cardiovascular diseases. Conventional methods such as examinations in healthcare facilities have limitations in terms of monitoring frequency, accessibility, and cost. Meanwhile, commercial wearable technologies such as smartwatches offer convenience but are generally limited in the number of parameters that can be monitored due to size and design constraints. The healthvest developed in this study is designed to overcome these limitations by offering a comfortable, stable, and accessible multiparameter monitoring solution. The system consists of electrocardiograph electrodes sewn into the inner part of the vest to ensure signal stability, along with a photoplethysmograph sensor unit embedded in a belt. electrocardiogram signals are analyzed for heart rate variability using discrete wavelet transform for R-R
interval extraction, followed by time-domain analysis and respiratory rate estimation through high-frequency components. Photoplethysmogram signals are used for blood oxygen level estimation with a machine learning approach based on random forest, and for blood glucose and blood pressure prediction using deep learning models based on convolutional neural networks and temporal convolutional networks equipped with a convolutional block attention module. All data are processed locally and transmitted wirelessly to the server with low latency (<1 ms), then displayed via a user-friendly interface. Experimental results show that the device produces stable signals and accurate vital sign estimations, with mean absolute errors (MAE) of 4.59 bpm for heart rate, 1.15 brpm for respiratory
rate, 1.13% for SpO2, 14.275 mg/dL for blood glucose, and 6.22 mmHg and 4.46 mmHg for systolic and diastolic blood pressure. This device offers a comprehensive and practical solution for daily health monitoring.

Item Type: Thesis (Other)
Uncontrolled Keywords: Rompi Kesehatan, Electrocardiogram, Photoplethysmogram, Pemantauan Kesehatan, Estimasi Tanda Vital
Subjects: R Medicine > R Medicine (General) > R856.2 Medical instruments and apparatus.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7872.F5 Filters (Electric)
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User: Aldo Pranata
Date Deposited: 04 Aug 2025 12:40
Last Modified: 04 Aug 2025 12:40
URI: http://repository.its.ac.id/id/eprint/126113

Available Versions of this Item

Actions (login required)

View Item View Item