Development of a wearable Pulse Wave Velocity estimation system using Force Sensing Resistor

Cardiovascular disease (CVD) is a class of diseases involving the heart or blood vessels and is the leading cause of death worldwide. It is estimated that around 18 million people die from cardiovascular disease each year, and this number corresponds to around 32% of global annual deaths. The chance of premature mortality due to CVD can be reduced by early disease diagnosis and analysis of several predictive parameters such as arterial stiffness, which is a key factor in cardiovascular physiology and is strongly related to Pulse Wave Velocity (PWV). PWV represents the rate at which the blood pressure pulse propagates through the circulatory system and is a measure of arterial stiffness. The examination is done non-invasively between two sensing sites, usually the carotid site and femoral site, and PWV is calculated as the ratio of distance and the time needed by the pulse to propagate between these two sites. Nowadays, PWV estimation is done using expensive technologies such as applanation tonometry and ultrasound, limiting their use in a clinical environment. Furthermore, some operational expertise is required to use applanation tonometry, the most widely used technology in this field, with the potential to influence the measurement as these devices are not wearable. This thesis aims to develop a clinical low-cost and wearable device for Pulse Wave Velocity estimation. The transducer used in this study is a Force Sensing Resistor (FSR), which is a piezoresistive force sensor that can detect the small force variations caused by artery displacement due to the pulse wave transit. The sensor response depends on the variation of the electrical resistance related to the applied force on it. Since FSR is a low-cost, flexible, and ultra-thin sensor and can interface with human skin, it is suitable for use in a wearable application such as the one discussed in this thesis. The characterization of the various FSR product types available on the market and the selection of the best type of sensor for the application comprised the first step of the presented work. A read-out circuit that can pick up and amplify the signal of interest has been studied and implemented around the chosen sensor. Subsequently, a Printed Circuit Board (PCB) has been realized to significantly reduce the dimension of the device. The PCB, connected to the sensor and encapsulated in a specifically developed support, interfaces with the Discovery Kit produced by STMicroelectronics, for which a specific firmware has been developed that enables the acquisition and transmission of pulse wave data to the PC. On the PC side, a Graphical User Interface (GUI) has been developed to visualize in real-time the signals taken from the carotid and femoral sites in a user-friendly way and with the possibility of saving data for the subsequent analysis carried out on Matlab, implementing the intersecting tangent method for the PWV extraction. Finally, the system has been tested on 14 volunteers at “A.O.U. Città della Salute e della Scienza di Torino” and statistical analysis has been conducted to compare the obtained results with those provided by SphygmoCor, the gold standard for PWV estimation. The results obtained showed how the proposed device’s performance is consistent with the gold standard as there is a small offset error and a strong correlation that is noticeable with a high coefficient of determination.