Non-Invasive Arterial Parameter Detection via Neural Network and Windkessel Model

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for determining arterial compliance, conduction resistance, and blood flowability are hindered by measurement uncertainties, limited reproducibility, and the need for invasive procedures, with pulse contour-based methods often failing to clearly separate these parameters due to superimposed waves and tissue interference.

Innovation Solution

A method using a trained neural network to analyze pulse curves, combined with an electro-hydraulic artery model, allows for non-invasive determination of arterial compliance, line resistance, and blood inertia by simulating the arterial tree and applying correction factors to improve measurement accuracy and reproducibility.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If pulse contour analysis is used to determine arterial parameters, then non-invasive measurement is achieved, but measurement precision deteriorates due to superimposed waves and tissue interference

Engineering Contradiction:
Improvetissue interferenceVSAvoidparameter determination accuracy
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The arterial system is segmented into multiple sections (proximal, middle, distal) with different Windkessel models applied to each segment. This allows separation of wave reflections from different anatomical locations, enabling precise determination of local arterial parameters despite superimposed waves in the composite pulse waveform.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

A multi-section Windkessel model serves as an intermediary framework that mathematically separates the contributions of different arterial segments to the observed pulse waveform. This mediator model enables extraction of precise local parameters (compliance, resistance, inertia) from the composite signal contaminated by tissue interference and wave superposition.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If characteristic points of pulse waveform are used to determine surrogate parameters, then non-invasive measurement is possible, but measurement precision deteriorates due to difficulty in clearly determining position and amplitude

Engineering Contradiction:
Improvenon-invasive measurementVSAvoidposition and amplitude determination
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The mechanical/visual method of identifying characteristic points on pulse waveforms is replaced with a mathematical model-based approach using multi-section Windkessel models. This substitution transforms the problem from subjective visual estimation to objective parameter extraction through model fitting, significantly improving precision of position and amplitude determination.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If arterial parameters are measured using traditional methods, then measurement can be performed, but reliability deteriorates due to measurement uncertainties and limited reproducibility

Engineering Contradiction:
Improvemeasurement capabilityVSAvoidreproducibility
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The measurement system incorporates feedback through iterative optimization where the multi-section Windkessel model parameters are adjusted to minimize the difference between simulated and measured pulse waveforms. This feedback mechanism ensures reliable and reproducible determination of arterial parameters by continuously refining the model fit to the actual physiological data.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3445233B1Method for detecting arterial parameters of a human being and device for carrying out said method
Publication Date: 2024.11.06 WSH ENG SERVICES GBR VERTRETUNGSBERECHTIGTER GESELLSCHADTER HIERONYMI ANDREAS
  • EP3445233B1 patent drawingFigure 1
  • EP3445233B1 patent drawingFigure 2
  • EP3445233B1 patent drawingFigure 3

AI summary

In order to detect arterial parameters, an averaged pulse curve of a human being is fed to a calculating unit. Said calculating unit adapts the parameters of a simulation model, a so-called analogue, in such a manner that the averaged pulse curve corresponds to the pulse curve of the simulation model. The arterial parameters output by the analogue can be displayed in a radar chart.