A continuous blood pressure monitoring method based on millimeter wave radar and multi-scale attention network
By using global normalization and multi-scale attention network to process radar signals, the problems of low blood pressure estimation accuracy and noise interference in existing technologies are solved, and high-precision continuous blood pressure monitoring is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- UNIV OF ELECTRONICS SCI & TECH OF CHINA
- Filing Date
- 2026-04-22
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies, when using deep learning models to process radar signals, cannot fully preserve the absolute physical properties and physiological characteristics of the signals, resulting in low accuracy in blood pressure estimation and susceptibility to noise interference.
Global maximum-minimum normalization is used to process radar pulse wave signals. Combined with a multi-scale attention network and an attention mechanism that integrates pulse pressure difference priors, multi-scale morphological features are extracted to enhance physiological feature extraction and noise resistance.
It achieves high-precision continuous blood pressure waveform reconstruction, with estimation errors of systolic and diastolic blood pressure as low as 4.54 mmHg and 2.82 mmHg, respectively, which has high clinical reference value.
Smart Images
Figure CN122398243A_ABST