A comprehensive multi-parametric PPG sensor platform targeting an in-depth understanding of the impact of the contact force
 
A comprehensive multi-parametric PPG sensor platform targeting an in-depth understanding of the impact of the contact force 
 
 
Abstract 

Photoplethysmography (PPG) has evolved into a versatile, non-invasive method for continuous cardio-vascular monitoring over the past two decades. Despite its widespread use for assessing cardiovascularhealth and monitoring vital signs, PPG{\textquoteright}s clinical utility is compromised by its sensitivity to factors like thecontact force (CF) exerted by the sensor and individual physiological variations. This work introduces anovel Multi-parametric Sensor System designed explicitly for multi-wavelength PPG (MW-PPG) signalcharacterization, utilizing a unique Multi-modal MW-PPG dataset collected in vivo. The research focuseson the dynamic responses of PPG signals to variations in CF, revealing that blood vessels closest to theskin are most affected, with their recovery dependent on the viscoelastic properties of the tissue andspecific anatomical features like bone structure in the fingers. The amplitude of the AC component of thesignals increases with CF as blood vessel walls compress but decreases as CF rises and arteries begin toocclude, resulting in complex waveform changes that vary with both wavelength and tissue properties. TheDC component strongly correlates with CF levels, indicating its potential use as an indirect measure ofCF. The study also demonstrates that the response of MW-PPG signals to CF is wavelength-dependent,with shorter wavelengths affected first due to shallower penetration. A distinct hysteresis pattern at the631 nm wavelength was identified, and tonometric signals aligned with deeper wavelength PPG signals,challenging existing assumptions about their relationship. Furthermore, the research establishes an optimalrange of normalized CF between 0.2 and 0.6 to ensure signal stability, emphasizing that signal qualityindices (SQIs), such as skewness and kurtosis, remain consistent. Normalized CF values over absolutevalues are recommended. This adjustment could significantly refine measurement precision, accommodateinterindividual variability, and enhance the consistency and reliability of signal quality across diversepopulations. Additionally, it facilitates more uniform comparisons between studies, improving the re-producibility of findings in the field. In light of the findings from this research, robust methods for PPGsignal acquisition are proposed, addressing factors frequently overlooked in the current state of the art.These methods include developing innovative devices, measurement techniques, and a novel calibrationscheme. A preliminary version of a measurement device designed for concept testing and validation ofthese advancements has also been developed.