Millimeter wave radar human body tumble detection method based on multi-feature fusion
Through multi-feature fusion methods, adaptive clustering, multipath filtering, deep time series learning and multimodal information are used to solve the problems of high false alarm rate and poor real-time performance of millimeter-wave radar in human fall detection in complex home environments, and achieve high-accuracy and low-resource consumption fall detection.
Patent Information
- Application Number
- CN202511141685.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-15
- Publication Date
- 2025-09-16
- Estimated Expiration
- 2045-08-15
AI Technical Summary
Existing millimeter-wave radar human fall detection methods have difficulty distinguishing between falls and lying down in complex home environments, have a high false alarm rate, and have limited adaptability to environmental noise and interference. The high resource requirements of deep learning models lead to poor real-time performance.
A multi-feature fusion method is adopted, including data preprocessing, dynamic clustering, multipath false cluster filtering, time series tracking, deep time series model and posture data fusion. By combining adaptive neighborhood radius density clustering, Kalman filtering, bidirectional LSTM network and multimodal information of visual devices, the detection accuracy and real-time performance are improved.
It significantly reduces the false alarm and missed alarm rates, improves detection accuracy and real-time performance in complex environments, reduces the demand for hardware resources, adapts to different heights and fall methods, and achieves fast response and stable fall detection.