Sensorex: The Challenges for Engineering Implementation of Low-Cost Non-Invasive Pulse Oximeter Applicable to Diverse Patient Population

Two wavelengths transmittance non-invasive pulse oximetry is commonly used in clinical practice. However, in a need of quick blood oxygen diagnostic assessment, a low-cost pulse oximeter needs to operate at a wide variety of patients using the same fingertip sensor or, at least, using a sensor of identical model and manufacturer. The average intensity for both optical and near-infrared light transmitted through the fingertip can be higher or lower depending on the volume of the finger, skin colour, tissue composition, and thus varies for people of different ages, gender, morphology, or national origin. Therefore, the device must correctly calculate the pulse oximetry parameters for signals with different tissue transmittances. Also, for the different age groups, different types of fingertip sensors are usually used (e.q. adhesive, clips, silicone mittens, etc.), the difference in fastening of the sensor can potentially distort the measurement results. A SENSOREX hardware and software provides correct extraction of saturation and perfusion coefficient and also extracts a plethysmographic wave suitable for subsequent adaptive signal processing implemented in the device to eliminate the effects of natural patient movement. At the time of the publication, the SENSOREX is at the final stage of Russian Federation state registration as a medical device for professional and home care usage. The implemented software-hardware solution is suitable for different categories of patients and uses with Nellcor - compatible fingertip sensors of various way of attachments. During the development, contradictions in the stakeholders' requirements and regulatory documents were considered, which entailed difficulties in engineering implementation. The paper presents the found methods to overcome those challenges and discusses the further promising development of the software and hardware complex and subsequent signal processing.