Wearable HbA1c Sensing Using Multi-Wavelength PPG and ML

The patent segments the spectral measurement task into multiple discrete wavelength channels (at least three different wavelengths) that can be measured sequentially rather than simultaneously. This allows using simpler, smaller detectors for each wavelength band rather than requiring a full conventional spectrometer, thereby reducing device bulkiness while maintaining measurement capability through multi-wavelength discrimination of hemoglobin species.

The patent employs periodic action by sequentially switching between different wavelength measurement channels during a measurement cycle. Instead of using all wavelengths simultaneously as in conventional spectrometers, the system periodically activates different wavelength pairs (e.g., green and red channels) to collect spectral data over time, enabling accurate HbA1c measurement with simpler hardware.

The patent introduces an intermediary processing step involving a trained machine learning model that acts as a mediator between the raw multi-wavelength optical signals and the final HbA1c concentration value. This model processes the complex signals containing skin reflection and overlapping absorption artifacts, separating the relevant hemoglobin absorption information from interfering signals through pattern recognition trained on reference datasets.

The patent utilizes parameter changes by measuring at multiple discrete wavelength points across the visible spectrum and using these spectral parameter variations to differentiate between oxyhemoglobin, deoxyhemoglobin, and glycated hemoglobin. The system analyzes how absorption characteristics change with wavelength to resolve overlapping spectra and extract accurate HbA1c information despite interference.

The patent replaces the mechanical/invasive approach of electrochemical sensors requiring skin penetration with an optical measurement system that uses non-invasive light absorption through the skin. By substituting mechanical contact with optical fields, the system achieves glucose and HbA1c measurement without breaking the skin, dramatically improving user convenience and enabling continuous monitoring.

The patent implements universality by designing a single optical device capable of measuring multiple parameters including glucose concentration and glycated hemoglobin levels using the same hardware platform. The multi-wavelength optical system can simultaneously or sequentially extract information about different blood components, providing versatile health monitoring functionality without requiring multiple separate devices or invasive procedures.