Pengembangan Algoritma Deteksi Karakteristik Poin pada Sinyal Doppler Cardiogram untuk Analisa Tanda Vital secara Non-Kontak

Towidjojo, Revelyno Rellert (2025) Pengembangan Algoritma Deteksi Karakteristik Poin pada Sinyal Doppler Cardiogram untuk Analisa Tanda Vital secara Non-Kontak. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211052-Undergraduate _Thesis.pdf Restricted to Repository staff only Download (18MB) | Request a copy

Abstract

Pemantauan tanda vital penting untuk diagnosis, terutama dalam pemantauan jangka panjang. Namun, metode berbasis sensor kontak kurang nyaman untuk penggunaan jangka panjang. Doppler radar dapat digunakan sebagai alternatif dalam memantau tanda vital secara non-kontak dengan mendeteksi pergerakan dada yang dihasilkan oleh aktivitas mekanik jantung. Sinyal Doppler yang ditangkap menghasilkan sinyal Doppler Cardiogram (DCG) yang memiliki karakteristik poin temporal yang sesuai dengan sinyal Electrocardiogram (ECG). Namun, Sinyal pernapasan ikut terdeteksi sehingga dapat mengganggu akurasi sinyal DCG dalam melakukan analisa tanda vital. Selain itu, deteksi otomatis karakteristik sinyal jantung dari Doppler radar masih terkendala oleh tumpang tindihnya sinyal pernapasan. Penelitian ini mengintegrasikan arctangent demodulation untuk ekstraksi fase gerak dinding dada, diikuti oleh Successive Variational Mode Decomposition (SVMD) untuk memisahkan sinyal pernapasan dan sinyal DCG. Selanjutnya, sinyal DCG dimasukkan ke dalam algoritma deteksi karakteristik poin dengan memanfaatkan fitur topologi yang dapat mengklasifikasikan setiap Karakteristik Poin pada DCG berdasarkan posisinya secara berurutan. Pengukuran dilakukan menggunakan radar Continuous Wave (CW) 24GHz dengan perekaman ECG secara simultan terhadap 14 subjek. Metode algoritma yang diajukan pada penelitian menunjukkan hasil identifikasi karakteristik poin secara konsisten dengan rata-rata pengukuran heart rate yang tinggi dengan Root Mean Square Error (RMSE) sebesar 12.69 BPM. Namun, analisis beat-to-beat menghasilkan RMSE 0.136 s untuk interval C-C terhadap R-R yang mengindikasikan keterbatasan sistem. Meskipun demikian, algoritma ini mampu mengekstraksi waktu interval jantung lebih baik dengan rata-rata RMSE sebesar 0.046 s untuk interval B-C terhadap P-R yang memperlihatkan korelasi fisiologis yang kuat dari titik yang terdeteksi. Keberhasilan algoritma dalam mengidentifikasi Karakteristik Poin secara andal dalam sinyal yang bervariasi menunjukkan potensinya yang signifikan untuk pemantauan tanda vital secara non-kontak.
=================================================================================================================================
Vital sign monitoring, such as heart rate and respiration, is essential for diagnosis and long-term patient care. However, conventional contact-based methods like ECG are often uncomfortable and unsuitable for prolonged use. Doppler radar offers a non-contact alternative by detecting chest wall movements caused by cardiac mechanical activity. The captured Doppler signal produces a Doppler Cardiogram (DCG), which contains temporal characteristic points analogous to those in an Electrocardiogram (ECG). Nevertheless, respiratory signals are also detected, which can interfere with the accuracy of DCG-based vital sign analysis. Moreover, automatic detection of cardiac features from radar signals remains challenging due to the overlap between cardiac and respiratory components. This study integrates arctangent demodulation for chest wall motion phase extraction, followed by Successive Variational Mode Decomposition (SVMD) to separate the respiratory and DCG signal. Next, the DCG signal is fed into a characteristic point detection algorithm that utilizes topological features to classify each characteristic point on the DCG based on its sequential position. Measurements were performed using a 24 GHz Continuous Wave (CW) radar with simultaneous ECG recording on 10 subjects. The proposed algorithm in the study showed consistent characteristic point identification results with a high average heart rate measurement with a Root Mean Square Error (RMSE) of 12.69 BPM. However, the beat-to-beat analysis resulted in an RMSE of 0.136 s for the C-C interval against the R-R interval, indicating a limitation of the system. Nevertheless, the algorithm demonstrated better performance in extracting other cardiac time interval with an average RMSE of 0.046 s for the B-C interval to P-R interval, which shows a strong physiological correlation of the detected points. The algorithm's success in reliably identifying Characteristic Points in varying signals demonstrates its significant potential for non-contact monitoring of vital signs.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Doppler Radar ,Doppler Cardiogram, Karakteristik Poin, Successive Variational Mode Decomposition, Tanda Vital, Characteristics Point, Vital Sign.
Subjects:	R Medicine > R Medicine (General) > R856.2 Medical instruments and apparatus. R Medicine > RC Internal medicine > RC683.5.E5 Electrocardiography T Technology > TA Engineering (General). Civil engineering (General) > TA1573 Detectors. Sensors T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions:	Faculty of medicine and health (MEDICS) > Medical Technology > 11503-(S1) Undergraduate Thesis
Depositing User:	Revelyno Rellert Towidjojo
Date Deposited:	01 Aug 2025 01:08
Last Modified:	01 Aug 2025 01:08
URI:	http://repository.its.ac.id/id/eprint/122791

Actions (login required)

View Item