Oleg Stelia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Oleg Stelia has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- Iurii Krak
Taras Shevchenko National University of Kyiv
1 shared publication - Mykola Yefremov
Department of Theoretical Cybernetics, Taras Shevchenko National University of Kyiv, Kyiv, Ukraine, Ukraine
1 shared publication - Olexander Barmak
Department of Computer Science, Khmelnytskyi National University, Khmelnytskyi, Ukraine, Ukraine
1 shared publication
Latest Publications
- Open Access
- 19 October 2026
- Published by MDPI AG in Preprints.org
Resource-Efficient Edge Framework for Real-Time ECG Processing and R-Peak Detection on 8-Bit Microcontrollers
The expansion of the Internet of Medical Things (IoMT) has increased the demand for real-time electrocardiogram (ECG) monitoring on wearable devices. However, implementing signal processing on ultra-low-power edge nodes is constrained by hardware limitations, specifically the absence of a Floating-Point Unit (FPU) and restricted SRAM capacity. This paper presents a resource-efficient framework for real-time R-peak detection optimized for 8-bit microcontrollers. The core component is a piecewise-polynomial smoothing algorithm based on Hermite splines, integrated directly into the ADC sampling loop. To maintain signal integrity under motion artifacts, a cascading filtering architecture is employed, operating on fixed-size frame buffers with a peak memory footprint of approximately 2.3 KB. Detection performance was evaluated on physical hardware signals (AD8232/ATmega2560) recorded during rest and physical activity, while reference recordings from the MIT-BIH Arrhythmia Database served as a morphological reference. Statistical analysis using ensemble averaging and point-wise confidence intervals (96% confidence level, D = 2.05) confirmed that the filtering process suppresses quantization noise and myographic artifacts without distorting the physiological morphology of the ECG. Experimental evaluation demonstrated a sensitivity (SEN) of 94.67–99.83%, a positive predictive value (PPV) of 99.30–100%, and a mean F-score of 98.95%. Extracting and transmitting metadata (RR intervals) instead of raw signal samples reduces network traffic and overall power consumption in autonomous edge-based cardiac monitoring nodes.
- 19 October 2026
- Published by MDPI AG in Preprints.org
Open AccessResource-Efficient Edge Framework for Real-Time ECG Processing and R-Peak Detection on 8-Bit Microcontrollers
The expansion of the Internet of Medical Things (IoMT) has increased the demand for real-time electrocardiogram (ECG) monitoring on wearable devices. However, implementing signal processing on ultra-low-power edge nodes is constrained by hardware limitations, specifically the absence of a Floating-Point Unit (FPU) and restricted SRAM capacity. This paper presents a resource-efficient framework for real-time R-peak detection optimized for 8-bit microcontrollers. The core component is a piecewise-polynomial smoothing algorithm based on Hermite splines, integrated directly into the ADC sampling loop. To maintain signal integrity under motion artifacts, a cascading filtering architecture is employed, operating on fixed-size frame buffers with a peak memory footprint of approximately 2.3 KB. Detection performance was evaluated on physical hardware signals (AD8232/ATmega2560) recorded during rest and physical activity, while reference recordings from the MIT-BIH Arrhythmia Database served as a morphological reference. Statistical analysis using ensemble averaging and point-wise confidence intervals (96% confidence level, D = 2.05) confirmed that the filtering process suppresses quantization noise and myographic artifacts without distorting the physiological morphology of the ECG. Experimental evaluation demonstrated a sensitivity (SEN) of 94.67–99.83%, a positive predictive value (PPV) of 99.30–100%, and a mean F-score of 98.95%. Extracting and transmitting metadata (RR intervals) instead of raw signal samples reduces network traffic and overall power consumption in autonomous edge-based cardiac monitoring nodes.