An SpO2 Sensor Using Reconstruction-Free Sparse Sampling for 70% System Power Reduction

Abstract

Low arterial blood oxygenation (SpO2) is a measure of hypoxemia and a sign of problems relating to breathing and circulation. Progressive drop in arterial SpO2 can be an early indicator of severe disease in COVID-19 patients [1]. A hypoxic state can occur rapidly and without a patient's knowledge; therefore, early detection of SpO2 decline can be lifesaving. In other respiratory system diseases such as COPD and sleep apnea, continuously monitoring of SpO2 with a pulse oximeter can enable timely diagnosis of oxygen desaturation. SpO2 is measured with Photoplethysmography (PPG) that uses a photodetector (PD) to detect either the transmission or reflection of light from the surface of the skin at two different light wavelengths. Commercial fingertip SpO2 sensors are not designed for chronic wear and require user intervention to trigger measurements. Alternatively, wearable SpO2 recording devices in the form of watches and rings can operate in the background with minimal user intervention. However, continuous acquisition of SpO2 can present a significant power burden to a wearable device since high-power LEDs must be powered on for each sample, dominating the power dissipation of the sensor. We present a low-power pulse oximeter sensor IC that utilizes sparse sampling to reduce the overall power by 70%.