The influence of external factors on the accuracy of non-invasive measuring of oxygen in blood

In this paper we investigated a pulse oximetry-based method for mobile devices. This method obtains bio-signals related to blood pulsation in transparent parts of body. The most widely accepted field for use of this method is hospital care. In these cases a pulse oximeter is the best solution for the monitoring of emergency patients. A promising field for pulse oximetry is physical exercise. It only requires simple clips such as ear-clips, finger-clips, headbands etc. However this method presents some difficulties: weak signal, noise ratio, motion artefacts, low perfusion. We used a MAX30100 Oximeter and Heart Rate Sensor integrated circuit to obtain signals of blood pulse waves from red and infrared light emission diodes (LED). This device measures the oxygen saturation of a person's blood by placing an LED and a photodetector against the thin skin of a person's body, such as a fingertip, wrist or earlobe. The MAX30100 is a 14-pin surface mount integrated circuit that contains sensors for measuring a person's heart rate. It can also indirectly determine the oxygen saturation of a person's blood. The MAX30100 provides a complete pulse oximetry and heart rate measurement solution for medical monitors and wearable fitness devices. As each LED emits light into a person's finger, the integrated photodetector measures variations in light caused by changes in blood volume. An integrated 16-bit analog to digital converter (ADC) with programmable sample rate converts the photodetector output to a digital value. The MAX30100 filters out ambient light that can interfere with an accurate reading. Data are read through a serial I2C interface to computer for further processing. The LED current can be programmed from 0 to 50 mA with proper supply voltage. The LED pulse width can be programmed from 200 mu s to 1.6 ms to optimize measurement accuracy and power consumption based on use cases. The SpO(2) algorithm is relatively insensitive to the wavelength of the infrared LED, but the red LED's wavelength is critical to correct interpretation of the data. The temperature sensor data can be used to compensate errors with ambient temperature changes. During the experiments we registered oxygen concentration values under different conditions. With low muscular activity we observed high stability and repeatability of measuring values under various exterior conditions. However, with high muscular activity there were various artefacts in the gauged signals that led to contortion of effects. We identified the boundaries of the validity of measuring and propose the use of an adapted filter in order to distinguish pulse waves from optical signals more reliably. These devices can be applied in fitness training, medical monitoring and used as wearable devices.