Multi-modal human body sensing data fusion analysis and risk pre-judgment method and system for intelligent medical care

By constructing a multi-source acquisition terminal network, timestamp calibration, and cross-modal attention fusion network, and dynamically updating the personalized analysis model, the problems of spatiotemporal alignment and coupling of multimodal data are solved, enabling real-time monitoring and personalized intervention of the health status of the elderly, and improving the accuracy and practicality of risk prediction in the smart medical and elderly care system.

CN122117381APending Publication Date: 2026-05-29GUANGZHOU INST OF RAILWAY TECH
0 Cites 1 Cited by

Patent Information

Application Number
CN202610198721.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-11
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

In existing smart healthcare systems, multimodal data lacks effective spatiotemporal alignment and deep correlation mining, making it impossible to accurately capture the coupling relationship between physiological changes, behavioral patterns and environmental factors. The analysis models are static and rigid, making it difficult to adapt to the dynamic evolution of the individual health status of the elderly. Risk judgment relies on ex-post threshold triggering, resulting in delayed early warnings and a high false alarm rate, failing to achieve real-time monitoring, early warning and precise intervention.

Method used

A multi-source acquisition terminal network is constructed, and spatiotemporal registration of multimodal data is achieved through a timestamp calibration algorithm. Uncertainty quantification processing and cross-modal attention fusion network are introduced to dynamically update the personalized analysis model and generate personalized intervention suggestions by combining a scenario-based rule base.

Benefits of technology

It enables dynamic perception and intelligent analysis of the health status of the elderly, improves the accuracy of health risk prediction and the real-time nature of personalized intervention, breaks through the problems of model rigidity and early warning lag in traditional methods, realizes the transformation from passive response to proactive early warning, and improves the overall efficiency of the smart medical and elderly care system.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a multi-modal human body sensing data fusion analysis and risk pre-judgment method and system for intelligent medical care, and the method comprises the steps: constructing a multi-source collection terminal network to synchronously collect physiological, behavior and environment data, and achieving the time-space registration through high-precision timestamp calibration and UWB space positioning; performing multi-level cleaning and uncertainty quantization on the data to generate preprocessed data with confidence; respectively extracting depth features by using a convolutional neural network, a graph neural network and the like, and dynamically fusing the depth features into a comprehensive feature vector through a cross-modal attention mechanism; inputting the vector into a personalized analysis model which is continuously and dynamically updated based on user feedback, and pre-judging and grading the early probability of the health risk by adopting a time sequence prediction algorithm; and finally, in combination with the real-time position of the user and the scenarized rule base, generating and outputting accurate personalized intervention suggestions. According to the invention, closed loop of perception, analysis, pre-judgment and intervention is realized, and the accuracy and practicability of intelligent medical care health monitoring are improved.
Need to check novelty before this filing date? Find Prior Art