Heart Beat Model for Reliable Detection Amid Movement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional heart rate detection methods using signal processing algorithms are unreliable when the monitored person makes movements, as these movements distort the signal and render the algorithms ineffective.
Innovation Solution
A method and apparatus that utilize a heart beat model created from signals measuring mechanical or acoustic activity of the cardiac system, updated using machine learning techniques, to detect heart beats even in the presence of movement, by incorporating respiratory cycle phase information and canceling interference from adjacent beats through inverse modeling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional signal processing algorithms are used for heart rate detection, then the detection is simple and fast, but the reliability deteriorates when the monitored person makes movements
Solution Approach 1:
The patent applies dynamics by making the heart beat model adaptive and updateable through machine learning techniques. The system continuously learns from new data to improve its ability to detect heart beats under varying conditions including movement, transforming a static algorithm into a dynamic, self-improving system that adapts to individual physiological variations and movement patterns.
Solution Approach 2:
The patent changes parameters by incorporating multiple features beyond simple peak detection, including heart beat shape information, heart beat interval information, and respiratory cycle phase information. These parameter changes enable the system to distinguish between signal variations caused by movement and actual heart beat patterns, significantly improving reliability during movement.
2Measurement precision
If the person stays still for accurate detection, then the signal quality is good, but the practical usability deteriorates
Solution Approach 1:
The patent converts the harmful effect of movement into a beneficial feature by training the machine learning model to recognize and compensate for movement-induced signal distortions. The system learns from examples containing movement artifacts, transforming what was previously a source of error into training data that improves the model's ability to handle real-world conditions where users cannot remain perfectly still.
3Measurement precision
If machine learning techniques are used to update the heart beat model, then the detection accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent applies partial action by selectively using machine learning techniques only when beneficial, rather than applying complex algorithms to all data continuously. The system can operate with simpler methods when conditions permit and switch to or update with machine learning models when improved accuracy is needed, balancing computational complexity with detection accuracy based on actual requirements.
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
The solution enhances the reliability of heart rate detection by accurately identifying heart beats and heart rate even when the person is moving, improving the precision of the detection process.
Implementation Method 1
In certain embodiments, the signal is a ballistocardiography signal.
Implementation Method 2
updating the heart beat model by training the model using a machine learning technique, such as supervised learning.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method that includes retrieving a signal that measures an activity of a cardiac system. A heart beat model is created based on the retrieved signal, and heart beats are detected using the heart beat model. The model can be trained using amachine learning technique and also respiratory cycle phase information can be used in training the model. The invention also relates to a related algorithm, apparatus and computer program.