Adaptive Heartbeat Detection with Noise-Signal Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cardiac monitoring technologies face challenges in accurately detecting heartbeats due to noise interference, particularly when using exogenous signals like accelerometers, as existing methods struggle to differentiate between noise and heartbeats, leading to potential loss of valuable information and ineffective noise cancellation.
Innovation Solution
The method involves detecting candidate peaks in cardiac signals, correlating them with peaks from another signal like an accelerometer, and adjusting peak likelihood based on temporal similarity, removing noise peaks if correlated, and generating heartbeat sequences from remaining peaks using temporal regularity and prominence measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If adaptive cancellation and wavelet coefficient removal are used to remove noise from cardiac signals, then noise reduction is improved, but valuable heartbeat information may be lost
Solution Approach 1:
The patent implements dynamic adjustment of peak likelihood scores based on temporal correlation with accelerometer peaks. Instead of static noise removal, the system continuously adapts the likelihood scoring by comparing candidate heartbeat peaks with accelerometer signal peaks, adjusting scores in real-time based on whether peaks occur during high-motion periods. This dynamic approach preserves heartbeat information while removing noise.
Solution Approach 2:
The system changes the parameter of peak likelihood scoring by introducing a correlation-based adjustment mechanism. When a candidate peak's timing correlates with accelerometer peaks indicating motion artifacts, its likelihood score is reduced. This parameter change allows selective suppression of noise peaks while maintaining true heartbeat peaks, resolving the contradiction between noise removal and information preservation.
2Measurement precision
If fixed delay correction is applied to align peaks between ECG and other cardiac signals, then temporal alignment is improved, but accuracy deteriorates in noisy conditions
Solution Approach 1:
The patent implements feedback by using the correlation between ECG peaks and accelerometer peaks to adjust the likelihood scoring of candidate heartbeats. The system continuously monitors the accelerometer signal and feeds this information back into the heartbeat detection process, adjusting peak likelihood scores based on whether peaks occur during motion periods. This feedback loop improves reliability in noisy conditions.
Solution Approach 2:
The system performs preliminary action by pre-identifying periods of high motion through accelerometer analysis before conducting final heartbeat detection. By anticipating when motion artifacts are likely to occur based on accelerometer data, the system can pre-adjust its peak detection criteria, improving reliability by being prepared for noisy conditions before they affect the ECG signal.
3Measurement precision
If combinatorial optimization is performed on a large number of candidate peaks, then heartbeat sequence detection is improved in noisy signals, but computational complexity increases
Solution Approach 1:
The patent applies partial action by focusing computational resources on evaluating only the most promising candidate peaks. Instead of performing exhaustive combinatorial optimization on all detected peaks, the system uses preliminary filtering based on peak prominence and temporal correlation with accelerometer data to identify a subset of high-probability candidates, then performs detailed sequence optimization only on these selected peaks, reducing computational complexity while maintaining detection accuracy.
Data Source
AI summary
A method for detecting heart beats is disclosed. A plurality of sensors are configured to receive a cardiac signal and another cardiac signal or a signal correlated with a noise source. A processor is configured to detect candidate peaks in a cardiac signal and select a subset of the candidate peaks for temporal correlation with features, such as peaks, in another cardiac signal or noise correlated signal. This relationship is quantified by a correlation measure. The correlation measure, in turn, influences the likelihood that a particular peak or sequence corresponds to a heartbeat. Candidate peaks that were not part of the correlation process may then be added to a sequence or sequences associated with the peaks subject to the correlation analysis. Sequences are scored according to quality and a final sequence is selected as possible heartbeats.