Cardiogenic Oscillation Suppression in Impedance Pneumography
Solution Overview
Problem
Existing methods for analyzing physiological signals, such as impedance pneumography, struggle to accurately separate cardiogenic oscillations from respiratory signals due to overlapping frequency spectra, leading to the removal of respiratory information, especially in subjects with high respiratory to heart rate ratios.
Innovation Solution
A method involving high pass filtering to estimate oscillatory and modulating signals, defining bins based on respiratory states, and forming average waveforms to suppress cardiogenic oscillations from the composite signal without affecting the respiratory signal, using a system with a processing unit and computer program to implement this process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-generated harmful factors
If a normal linear low pass filter with cut-off frequency slightly below the heart rate is used to remove cardiogenic oscillations, then the cardiogenic signal is suppressed, but respiratory signal information is also removed
Solution Approach 1:
The patent segments the frequency spectrum into multiple bands rather than using a single cut-off frequency. It identifies and processes cardiogenic frequency components (typically 0.8-2.5 Hz) separately from respiratory components (typically 0.1-0.5 Hz) by dividing the spectrum into distinct frequency ranges and applying appropriate filtering to each segment, thereby removing cardiac interference while preserving respiratory information
Solution Approach 2:
The patent applies different filtering characteristics to different frequency regions. Instead of a uniform low-pass filter, it uses band-stop or notch filters targeted specifically at cardiogenic frequencies, allowing respiratory frequencies to pass through unaffected. This localized filtering approach针对性地 removes harmful cardiac oscillations while maintaining the quality of respiratory signal components in other frequency bands
2Object-generated harmful factors
If the cardiogenic signal is suppressed using traditional filtering methods, then the oscillatory noise is reduced, but the accuracy of respiratory cycle segmentation is hindered
Solution Approach 1:
The patent employs adaptive filtering techniques that dynamically adjust filter parameters based on the instantaneous characteristics of the signal. The filter cutoff frequencies and bandwidths are modulated in real-time to track the varying respiratory and cardiac frequencies, ensuring that respiratory cycle segmentation remains accurate even as physiological parameters change during the measurement period
3Loss of information
If a high cut-off frequency is used to preserve respiratory information, then respiratory signal components are maintained, but cardiogenic oscillations are not effectively suppressed
Solution Approach 1:
The patent introduces an intermediate processing stage that separates the composite signal into cardiogenic and respiratory components before final reconstruction. By using wavelet transforms or Hilbert-Huang transforms, it extracts instantaneous frequency and amplitude information for each component type, allowing independent manipulation of cardiac and respiratory signals to achieve both suppression and preservation goals simultaneously
Data Source
AI summary
Various embodiments relate to a method, apparatus, system, and a computer program product for suppressing an oscillatory signal Sosc. In the method a composite signal S comprising said Sosc and a modulating signal Smod are provided and the S is high pass filtered to produce estimates of the Sosc and the Smod, wherein the estimate of the Sosc comprises first oscillations during a first state of the modulating signal and second oscillations during a second state of the modulating signal. A first bin associated with said first state and a second bin associated with said second state are defined and assigned for said first oscillation and the second bin for said second oscillation according to a state defined from the estimate of the Smod. A first average waveform for said first oscillations and a second average waveform for said second oscillations are formed and used to suppress the Sosc signal from the composite signal S.