Respiration Rate Detection Using Histogram Weighting for Motion Artifact Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing devices are not reliable in determining the respiration rate of children due to their fast and irregular breathing patterns, as well as motion artifacts that interfere with sensor signals.
Innovation Solution
A method and apparatus that analyze breathing-related features from sensor signals, such as accelerometers or heart rate sensors, by forming histograms and applying weighting to emphasize the most frequent breath periods, thereby improving the accuracy of respiration rate measurement by considering multiple breathing rates and filtering out motion artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing devices use simple sensor signals to measure respiration rate, then the device complexity is low, but the measurement precision deteriorates for children due to fast and irregular breathing patterns
Solution Approach 1:
The patent segments the breathing-related signal into multiple candidate breath periods by identifying zero-crossings and peaks/troughs. Each candidate breath period is analyzed separately to determine candidate respiration rates, allowing the system to handle irregular breathing patterns by processing individual breath events rather than assuming regular intervals
Solution Approach 2:
The patent transforms the raw sensor signal into multiple derived parameters including candidate breath periods, candidate respiration rates, and histogram distributions. By changing from a single raw signal to multiple processed parameters, the system can accurately capture fast and irregular breathing patterns characteristic of children's respiration
2Measurement precision
If the device uses manual counting method as golden reference, then the measurement precision is high, but the productivity is low due to requiring full minute counting by trained clinician
Solution Approach 1:
The patent replaces the mechanical manual counting method with an automated electronic signal processing system. The processing unit automatically identifies zero-crossings, peaks, and troughs in sensor signals to determine candidate breath periods and respiration rates, eliminating the need for trained clinicians to manually count breaths while maintaining measurement accuracy
Solution Approach 2:
The system performs self-service by automatically processing sensor signals to generate respiration rate measurements without requiring human intervention. The processing unit autonomously analyzes candidate breath periods, forms histograms, determines weighted average centers, and outputs final respiration rates, enabling rapid continuous measurements
3Reliability
If the device processes all candidate breath periods equally, then the calculation is simple, but the reliability deteriorates due to motion artifacts interfering with sensor signals
Solution Approach 1:
The patent implements feedback by forming a histogram from candidate respiration rates and using the histogram distribution to identify and weight reliable measurements. The weighted average center calculation uses feedback from the histogram shape to determine which candidate rates are most representative, filtering out motion artifact contamination through statistical analysis of the rate distribution
Solution Approach 2:
The system changes from processing single candidate rates to analyzing the distribution of multiple candidate rates through histograms. By examining the statistical distribution and applying weighted averaging based on histogram characteristics, the system can distinguish true respiration rates from motion artifacts that appear as outliers in the distribution
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to an aspect, there is provided a method of determining the respiration rate of a subject, the method comprising obtaining a signal from a sensor that is worn or carried by the subject; analyzing the signal to determine a plurality of values for a breathing-related feature; forming a histogram from the plurality of values for the breathing- related feature, the histogram comprising a plurality of groups, with each group having an associated count that is the number of occurrences of a value or values for the breathing-related feature corresponding to the group; applying a weighting to the count associated with each group to form weighted counts; and determining the respiration rate from a mean of the histogram with the weighted counts.