Continuous blood pressure monitoring allows for detecting the early onset of cardiovascular disease and assessing personal health status. Conventional piezoelectric blood pressure monitoring techniques have the ability to sense biosignals due to their good dynamic responses but have significant drawbacks in terms of power consumption, which limits the operation of blood pressure sensors. Although piezoelectric materials can be used to enhance the self-powered blood pressure sensor responses, the structure of the piezoelectric element can be modified to achieve a higher output voltage. Here, a structural study on piezoelectric metamaterials in blood pressure sensors is demonstrated, and output voltages are computed and compared to other architectures. Next, a Bayesian optimization framework is defined to get the optimal design according to the metamaterial design space. Machine learning algorithms were used for applying regression models to a simulated dataset, and a 2D map was visualized for key parameters. Finally, a time-dependent blood pressure was applied to the inner surface of an artery vessel inside a 3D tissue skin model to compare the output voltage for different metamaterials. Results revealed that all types of metamaterials can generate a higher electric potential in comparison to normal square-shaped piezoelectric elements. Bayesian optimization showed that honeycomb metamaterials had the optimal performance in generating output voltage, which was validated according to regression model analysis resulting from machine learning algorithms. The simulation of time-dependent blood pressure in a 3D skin tissue model revealed that the design suggested by the Bayesian optimization process can generate an electric potential more than two times greater than that of a conventional square-shaped piezoelectric element.
Read full abstract