| dc.description.abstract | Cardiovascular diseases are a leading contributor of health problems all over the world and are the second leading cause of death. They are also the cause of significant economic burden, costing billions of dollars in healthcare every year. With an aging population, the strain on the healthcare system, both in terms of costs and care provision, is expected to worsen.
Frequent cardiac assessment can provide essential information towards diagnosis, monitoring, and treatment, which can mitigate symptoms and improve health outcomes for people with conditions such as heart failure. This has led to increasing interest in cardiac assessment at home. Additionally, for some populations like people with limited mobility and older adults, long term vitals monitoring at a clinical setting is not feasible, making at-home monitoring more viable and economical. Most devices available for cardiac monitoring at home are wearables. While wearable technology can be accurate, it requires compliance and maintenance, which is not an ideal solution for all populations. For example, people who are not comfortable using wearables or people with a cognitive impairment may not want or be able to use wearables, which could exclude these user types from at home monitoring. Keeping these factors under consideration, the past decade has seen an increased interest in the development of technologies for Ambient Assisted Living (i.e., smart technologies integrated into a user's environment). These technologies have the potential for ongoing health monitoring in an unobtrusive manner.
This thesis presents research into the development of a smart seat cushion for heart rate monitoring. The cushion is able to calculate the heart rate of a person seated on it by acquiring their Ballistocardiogram (BCG). BCG is a cardiovascular signal corresponding to the displacement of the body in response to the heart pumping blood at every heartbeat. The prototype seat cushion has load cells embedded inside it that sense the micromovements of the body and translate it to an electrical signal. An analog signal conditioning circuit amplifies and filters this signal to enhance the components corresponding to BCG before it is converted to digital form. A pilot study was conducted with twenty participants to acquire BCG in real-world scenarios: 1) sitting still, 2) reading, 3) using a computer, 4) watching TV, and 5) having a conversation. Heart rate was calculated using a novel algorithm based on Continuous Wavelet Transform by detecting the largest peaks (referred to as the J-peaks) in the BCG. Excluding three outliers, the algorithm is able to achieve an overall accuracy of 94.6% compared to gold standard Electrocardiography (ECG). This accuracy is observed to be as good as or better than those of existing wearable heart rate monitors.
The seat cushion developed in this thesis research can serve as a portable solution for cardiac monitoring and can integrate into an ambient health monitoring system, offering continued monitoring of heart rate while requiring no perceived effort to operate it. Future work includes exploring different sensor configurations, machine learning based approaches for improving J-peaks detection, and real-time monitoring of heart rate. | en |
