Area of Applications: Wearable Healthcare & Telemedicine.
Engineering Expertise Involved:
The main goal is to develop a cost-effective wearable system for constant real-time measurement, processing and wireless transmission of biomedical data such as skin Temperature, Heart Rate (HR), Respiration Rate (RR) and Electrocardiogram (ECG) with required accuracy while withstanding acceptable degree of patient motion. The system will be designed with considerations of special care of COVID-19 patients providing remote monitoring of their most critical physiological parameters, but can be also widely used for any other telemedical purposes.
The unit will be positioning on upper arm and providing data sampling, analog signal pre-conditioning and normalization, and wireless transfer via Bluetooth Low Energy (BLE) radio link to the base station and networked PC for further post processing and analysis. For immediate detection of abnormal data patterns (vital signs), some initial data analysis will be performed in the host application at the base station, while more detailed evaluation can be executed on a server data management platform due to required higher computation resources.
The required accuracy will be achieved by applying Technesys’ proprietary signal processing and analysis algorithms employing combination of data pre-processing, signal morphology modeling and identification, signal pattern templating, data synchronization and alignment, peak detection, data segmentation, SNR estimation, and other data analysis techniques facilitating the design.
Analog Front-End (AFE) electronics is designed to sample and process ultra-low electrical potentials from skin for measuring biomedical data such as ECG and PPG. The signal pre-conditioning techniques involve signal denoising (i.e. active/passive versions of high-pass filters (HPF) and low-pass filters (LPF), baseline wandering stabilization, signal differentiating, automatic gain control, etc.).
This work also involves developing signal quality analysis SW algorithms comprising of (i) compensation of baseline wander artifacts, (ii) SNR estimation, (iii) Digital Signal Processing (DSP) – signal denoising based on classical filtering and adaptive median filtering followed by linear variational denoising, (iv) peak finder algorithm based on signal morphology templating, (v) pulse rate calculation, (vi) estimation of false positive/negative pulses. The developed algorithms have been verified for accuracy by conducting in-lab experimental comparison tests using medical grade devices as references.
The signal quality analysis has been performed for different motion conditions such as sitting, standing, walking, laying, arm roaming. Adaptive pre-conditioning algorithms have been developed for mitigating the motion artifacts.