Implantable Cardiac State Classification Using Low-Power Signal Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating cardiac signals in implantable devices face challenges such as high computational effort, power consumption, and reliance on training data quality, making them unsuitable for power-saving detection devices and prone to failure.
Innovation Solution
A method that subdivides cardiac signals into blocks of varying widths, calculates signal swing measures, and uses quotients to classify physiological and pathophysiological states without event-based segmentation, allowing for power-efficient and robust state differentiation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If event segmentation is used to evaluate cardiac signals, then the evaluation can be performed based on extracted events, but additional computational effort and power consumption are required
Solution Approach 1:
The patent extracts only the essential features needed for cardiac state classification directly from the signal segments, avoiding the need for complex event segmentation. By taking out only the necessary computational steps (signal segment extraction, feature calculation, classification), the method maintains evaluation reliability while removing unnecessary computational overhead that consumes power.
Solution Approach 2:
The patent applies segmentation by dividing the cardiac signal into fixed-time segments and further into blocks for feature calculation. This segmentation approach replaces complex event-based segmentation with simpler temporal segmentation, reducing computational effort while maintaining the ability to detect cardiac states accurately.
2Productivity
If computationally intensive signal transformations are used in block-based signal evaluation, then signal processing capability is improved, but power consumption increases making it unsuitable for implantable devices
Solution Approach 1:
The patent uses simple, computationally inexpensive operations (amplitude calculations, threshold comparisons, quotient formation) that can be executed quickly and discarded, replacing complex signal transformations. These simple operations consume minimal power while still providing sufficient signal processing capability for cardiac state classification in implantable devices.
Solution Approach 2:
The patent substitutes complex mathematical signal transformations with simpler arithmetic operations and logical comparisons. Instead of using intensive signal processing algorithms, the method uses basic computations (summing amplitudes, comparing to thresholds, forming quotients) that achieve the same classification goal with minimal power consumption.
3Measurement precision
If heuristic approaches with training data are used to differentiate cardiac states, then state differentiation capability is improved, but robustness depends on training data quality and additional computational resources are needed
Solution Approach 1:
The patent changes the approach from learning-based parameter adaptation to fixed threshold-based classification. By using predetermined thresholds and fixed classification rules that do not depend on training data quality, the method achieves robust state differentiation that is reliable across different patients and conditions without requiring extensive training datasets.
Solution Approach 2:
The classification system uses self-contained rules that do not require external training data or continuous learning. The predetermined thresholds and classification logic enable the system to differentiate cardiac states autonomously based on the signal characteristics themselves, making the system robust and reliable without external dependencies.
4Use of energy by moving object
If simple signal evaluation methods are used to reduce power consumption, then power efficiency is improved, but accuracy in differentiating physiological and pathophysiological states may be compromised
Solution Approach 1:
The patent applies local quality by calculating signal features in specific time blocks within segments and using localized amplitude comparisons. This localized analysis approach maintains classification accuracy by focusing computational effort on relevant signal portions while keeping overall power consumption low through selective rather than continuous complex processing.
Solution Approach 2:
The patent uses partial action by implementing only the essential steps needed for accurate classification (segmentation, amplitude calculation, threshold comparison, quotient formation) without performing unnecessary additional processing. This partial approach achieves sufficient accuracy for clinical decision-making while maintaining power efficiency in implantable devices.
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3D
AI summary
The invention relates to a device for detecting a signal from a human or animal organism which, during operation, carries out the following steps: detecting a signal (200, 610) from a human or animal organism in a time-dependent manner; subdividing a signal segment (201) into first blocks (210); determining a total number of the first blocks (210); determining a measure for a signal swing (220, 320, 420) in each of the first blocks (210); determining a number of first blocks (210) in which the measure for the signal swing (220, 320, 420) is less than a predeterminable first threshold value (330, 430); calculating a first quotient (q1) from the number of first blocks (210) in which the measure for the signal swing (220, 320, 420) is less than the first threshold value (330, 430) and the total number of first blocks (210); comparing the first quotient (q1) with a second threshold value (350, 450); classifying a state of the human or animal organism as physiological or as pathophysiological as a function of the previous comparison.