Wearable Doppler Sensor Action Recognition via Blood Flow Signal Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wearable devices face challenges in accurately recognizing user actions due to noise in blood flow information caused by body motion, which affects the accuracy of action and action part recognition.
Innovation Solution
The method and device acquire target blood flow information from specific parts of the body and use reference information to determine the action and action part by identifying differences in blood flow patterns, such as amplitude values or waveforms, to differentiate between normal and action-induced signals, thereby enhancing recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If blood flow information is used to recognize user actions, then action recognition capability is improved, but measurement precision deteriorates due to noise from body motion
Solution Approach 1:
The patent extracts and separates the action-related blood flow signal components from the motion-noise components by comparing target blood flow information with reference blood flow information. This extraction process isolates the useful action signals while removing the harmful motion noise, thereby improving measurement precision while maintaining action recognition capability.
Solution Approach 2:
The patent introduces reference blood flow information as an intermediary element to eliminate motion noise. By using the reference information (obtained from the same sensor during resting state) to compare against target blood flow information, the system creates a noise-cancellation mechanism that purifies the action detection signals.
2Adaptability or versatility
If Doppler measurement information is used to detect blood flow, then action detection capability is improved, but reliability deteriorates due to motion-induced noise
Solution Approach 1:
The patent converts the harmful motion noise into a beneficial differentiation feature. By comparing target blood flow information with reference information, the motion artifacts that were previously considered noise are transformed into useful discriminators that help identify action states, thereby improving both reliability and detection capability.
Solution Approach 2:
The system uses reference blood flow information as feedback to continuously compare against target measurements. This feedback mechanism allows the system to identify and filter out motion-induced errors, improving the reliability of blood flow measurements while maintaining action detection accuracy.
3Device complexity
If action recognition is performed without filtering noise, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent performs preliminary action by acquiring reference blood flow information during a resting state before action detection. This preliminary reference data is stored and then used to process subsequent target blood flow information, enabling noise filtering without adding complex real-time processing algorithms during action detection.
Solution Approach 2:
The patent changes the parameter approach by transforming raw blood flow signals into normalized difference values by comparing with reference information. This parameter transformation simplifies the signal processing while improving measurement precision, as the normalized differences directly represent action-induced changes rather than absolute signal values.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach improves the accuracy of action and action part recognition in wearable devices by effectively filtering out noise from motion-related blood flow information, allowing for more precise identification of user actions.
Implementation Method 1
acquiring target Doppler measurement information of the first part or a second part corresponding to the first part
Data Source
AI summary
The present application provides methods and devices for determining an action or an action part, and generally relates to the field of wearable devices. A method disclosed herein comprises: in response to detecting a motion of a first part of a body of a user, acquiring, using a photoelectric sensor, target Doppler measurement information of the first part or a second part corresponding to the first part; determining target velocity related information corresponding to the target Doppler measurement information; and determining the first part or the action according to the target velocity related information and reference information, wherein the target velocity related information comprises target blood flow velocity information or target blood flow information. The methods and devices provide a new scheme for recognizing an action and/or an action part.