An Intelligent Non-Invasive Blood Pressure Monitoring System Using Dual-Channel Separation and Asymmetric Gaussian Fitting
Year: 2026
Type: conference-paper
Abstract: Conventional portable electronic sphygmomanometers frequently struggle with weak signal extraction, pneumatic interference, and poor individual adaptability. To overcome these limitations, this paper presents an intelligent non-invasive blood pressure monitoring system based on the STM32 architecture. Regarding the hardware, an AC/DC dual-channel separate sampling architecture was employed to isolate and amplify weak oscillometric signals from the static pressure baseline, thereby overcoming the resolution bottleneck of traditional analog-to-digital converters. In terms of control strategy, a first-platform occlusion determination method based on mean absolute deviation and a segmented variable-time stepwise deflation logic were developed to optimize the balance between measurement precision and patient comfort. At the algorithmic level, a “peak-to-trough difference” method was utilized to reconstruct oscillation sequences, while an asymmetric Gaussian model was introduced for continuous pulse envelope reconstruction. Furthermore, a mean arterial pressure (MAP)-driven linear continuous mapping strategy was employed to calculate systolic and diastolic blood pressures. Experimental results demonstrate that the mean errors for systolic and diastolic pressures, compared to a standard mercury sphygmomanometer, are 1.2 mmHg (SD: 3.12 mmHg) and 1.6 mmHg (SD: 4.38 mmHg), respectively. The system fully complies with AAMI international standards, exhibiting high accuracy and stability.
Institution: Linyi University
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7223462425