A Wearable Device for Non-invasive Cardiac Monitoring

Wearable continuous cardiovascular and physical activity monitors are becoming an integral part of comprehensive health management programs aimed at reducing hospital readmission, improving adherence to home- and tele- rehabilitation programs and increasing patients’ compliance and engagement in their wellness management plans. Furthermore, earlier detection of a heart disease can then lead to earlier intervention and treatment. Motivated by this observation, this thesis addresses the problem of designing a wearable device performing data acquisition and processing so that the device can be used to monitor heart failure patients not only in the hospital but also during day-to-day living activities such as walking, exercising, showering and sleeping. This device will increase cost effectiveness of tele-rehabilitation and provide objective health information to promote reimbursable clinical services. The most sought after features of this class of devices include unobtrusiveness, accuracy, ease of use and patient comfort and medical cost savings. The Holter monitor is currently the most widely used device for outpatients, which usually records only the ECG trace and it does not typically have any real time communication; moreover, these class of devices cannot handle motion artifacts properly or the quality of recorded signal are degraded considerably in the presence of high level of exertion. The wearable device designed for this work, on the other hand, is able to record up to two leads ECG or one lead plus respiration, an embedded chest and an external finger probe photoplethysmogram(PPG). It has Bluetooth Low Energy (BLE) connectivity to send real-time data to a smartphone equipped with a graphical interface to display the readout and it is able to send the data recorded to a Cloud Server or by email. The capabilities of the designed monitoring device is complemented by Photoplethysmogram (PPG) inclusion with the purpose of measuring hemodynamic indexes and vital parameters that take into account also the peripheral circulation. In particular, a novel approach is developed to acquire the PPG from the chest which is usually associated with low dynamic range and low signal to noise ratio and therefore it requires more robust extraction algorithms. Along with the device and the graphical interface, a set of digital processing techniques have been implemented to extract vital information from the raw data such as heart rate, heart rate variability and Pulse Transit Time in order to ensure the integrity and validity of the data even during activities. The proposed methods rely heavily on signal processing tools such as linear filtering and processing, adaptive filtering as well as a decisional algorithm that utilizes wavelet delineator. The proposed algorithm increase accuracy of analysis features extraction during exertion or exercise and it achieves a high sensitivity with very low false positive observed.