Use this URL to cite or link to this record in EThOS:
Title: Advancement in wearable, long-term heart rate and rhythm monitoring
Author: Lynn, William David
ISNI:       0000 0004 6347 3761
Awarding Body: Ulster University
Current Institution: Ulster University
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Many devices have been produced for non-clinical rate monitoring as a statistical aid to fitness programmes. Although capable of rate monitoring, clinically the monitors have little prognostic capability, making them a poor choice as a medical tool. The preferred option is to monitor the electrical activity of the heart - The electrocardiogram (ECG) - opening up the possibility of rhythm analysis and morphological scrutiny. A clinical database of distal electrogram recordings was created in conjunction with the Craigavon Area Hospital Cardiac Research Department. Signal averaged ECG (SAECG) methods were then used to inspect electrograms recorded bilaterally in a pilot study and the evidence based outcome of which directed the research group to consider the left arm as a prime location for a potential long term cardiac monitor. The work also characterised the signal levels along the arms, showing rapid degradation of the ECG signal with respect to in-band noise. Empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD) and data fusion (DF) techniques were developed due to their ability to extract morphologically intact information from a dynamic data stream and their performance compared to the control SAECG reference method and clinically accepted denoising approach in high-resolution electrocardiography. EEMD was found to be a robust, low latency denoising technique, in comparison to SAECG performance; achieving signal to noise enhancement figures that improved on the SAECG control method, when used with far-field bipolar leads along the left arm ECG data. The EEMD denoising technique, based on a partial reconstruction of the original clinical data, when used with signals with a signal to noise ratio (SNR) greater than 30, produced results ranging within 8.9% (on average) of those produced by the SAECG control technique. As the SNR for the target signal reduced, the viability of the data driven techniques increased, with an average improvement over SAECG of 62.7% observed for the case of SNR of 1.9.An additional development to the EEMD, in the form of a data fusion between EEMD and frequency domain analysis enables 'automation' of the process for embedded application.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available