<![CDATA[SpO₂ is a critical physiological parameter in vital sign monitoring, particularly within Internet of Things (IoT) based healthcare systems that enable continuous and remote patient observation. However, the accuracy of SpO₂ measurements is often compromised by motion artifacts in photoplethysmograph (PPG) signals, which introduce amplitude distortion and frequency spreading that degrade oxygen saturation estimation. Although various signal processing approaches have been proposed, limited studies have evaluated digital filtering performance on real PPG signals using real-time embedded platforms with direct comparison of filter characteristics. This study aims to analyze the effectiveness of two Infinite Impulse Response (IIR) digital filters, Butterworth and Elliptic, in suppressing motion artifacts in SpO₂ signals processed on a microcontroller. Data were collected from ten healthy participants under three conditions: baseline (no movement), induced finger motion, and filtered signals. Signal quality improvement was assessed using Fast Fourier Transform (FFT), Power Spectral Density (PSD), and Signal-to-Noise Ratio (SNR) analyses. Results indicate that motion artifacts increase high-frequency components above 3 Hz and disrupt the morphological integrity of the PPG waveform. Applying a low-pass IIR filter with a 3 Hz cutoff frequency successfully restored the principal periodic components. The Butterworth filter produced a smoother spectral response with minimal phase distortion, while the Elliptic filter achieved a sharper roll-off with slight passband ripple. Quantitative evaluation demonstrated average SNR improvements of +0.905 dB (Butterworth) and +0.899 dB (Elliptic) using FFT , and +0.98 dB and +0.66 dB, respectively, using PSD. These findings demonstrate that computationally efficient IIR filtering can be reliably implemented in resource-constrained embedded platforms without compromising signal integrity. This approach enhances signal stability, reduces false desaturation alarms, supports scalable deployment in wearable and telemedicine applications, improving patient safety and system robustness in continuous remote health monitoring.]]>
