Date of Graduation
Doctor of Philosophy in Engineering (PhD)
Second Committee Member
Third Committee Member
Proper estimation of body fluid status for human or animal subjects has always been a challenging problem. Accurate and timely estimate of body fluid can prevent life threatening conditions under trauma and severe dehydration. The main objective of this research is the estimation, classification and detection of dehydration in human and animal subjects using peripheral blood pressure (PBP) signals. Peripheral venous pressure (PVP) and peripheral arterial pressure (PAP) signals have been investigated in this research. Both PVP and PAP signals are PBP signals. A dataset of PVP signals was collected using standard peripheral intravenous catheters from human subjects suffering from hypertrophic pyloric stenosis. Using this dataset, we successfully classified dehydrated subjects from hydrated subjects using regularized logistic regression on frequency domain data of the PVP signals. During the data acquisition process, the PVP signals was corrupted by noise and blood clot. So, we developed an unsupervised anomaly detection algorithm for PVP signals using hidden Markov model and Kalman filter. This anomaly detection algorithm removed the human bias in data-preprocessing. Another dataset of PAP and PVP signals was collected from pigs under anesthesia using the Millar catheter. We proposed a integral pulse frequency modulation (IPFM) based signal model for both PAP and PVP signals. The proposed model-synthesized signal is highly correlated with the experimental data. The model-synthesized signals also performs similar to experimental signals under classification tasks. We also examine the model estimated parameters both qualitatively and quantitatively. This model can also quantify the effect of respiratory rate on heart rate variability. Increasing doses of anesthesia has similar effect of getting hydrated from dehydration.
Hayat, M. (2023). Understanding Peripheral Blood Pressure Signals: A Statistical Learning Approach. Graduate Theses and Dissertations Retrieved from https://scholarworks.uark.edu/etd/4890