Method for detecting abnormal respiration using a photoplethysmography (PPG) signal

An abnormal respiration detection method comprising:

• sensing a photoplethysmography (PPG) signal from an optical signal reflected and received from a user's body;
• performing a band pass filtering for extracting a signal of a required band from the sensed PPG signal;
• performing a digital conversion on the filtered PPG signal;
• and
• deriving a respiration rate signal from the digital-converted PPG signal, deriving a plurality of respiration rate characteristic values through a time-axis analysis on the derived respiration rate signal, and detecting an abnormal respiration using the derived plurality of respiration rate characteristic values,
• wherein the plurality of respiration rate characteristic values include peak-to-valley values of the respiration rate signal expressed on the time axis and the number of peaks per unit time of the respiration rate signal,
• wherein the detecting an abnormal respiration is performed by a first analysis using the peak-to-valley values of the respiration rate signal and then a second analysis using the number of peaks per unit time of the respiration rate signal, the second analysis being successively performed if the first analysis indicates the detection of abnormal respiration,
• wherein the first analysis is performed according to Equation 1 by comparing a difference between a maximum peak-to-valley value of the peak-to-valley values during a current unit time and a previously-calculated moving average of the peak-to-valley values during a previous unit time with a first reference value,
• {
• 
• PPcurrent
• -
• PP
• ⁢
• average
• 
• PP
• ⁢
• average
• >
• 10
• ⁢
• %
• ,
• first
• ⁢
• ⁢
• detection
• ⁢
• ⁢
• of
• ⁢
• ⁢
• apnea
• 
• PPcurrent
• -
• PP
• ⁢
• average
• 
• PP
• ⁢
• average
• ≤
• 10
• ⁢
• %
• ,
• normal
• ⁢
• ⁢
• respiration
• [
• Equation
• ⁢
• ⁢
• 1
• ]
• where PPcurrent represents the maximum peak-to-valley value during the current unit time, and PPaverage represents the previously-calculated moving average of the peak-to-valley values, and
• wherein the second analysis is performed according to Equation 2 by comparing a difference between the number of peaks during the current unit time and a previously-calculated moving average value of the number of peaks during the previous unit time with a second reference value,
• {
• 
• PNs
• ,
• current
• -
• PNs
• ,
• average
• 
• PNs
• ,
• average
• >
• 40
• ⁢
• %
• ,
• second
• ⁢
• ⁢
• detection
• ⁢
• ⁢
• of
• ⁢
• ⁢
• apnea
• 
• PNs
• ,
• current
• -
• PNs
• ,
• average
• 
• PNs
• ,
• average
• ≤
• 40
• ⁢
• %
• ,
• normal
• ⁢
• ⁢
• respiration
• [
• Equation
• ⁢
• ⁢
• 2
• ]
• where PNcurrent represents the number of peaks during the current unit time, and PNaverage represents the moving average of the number of peaks during the previous unit time which has been previously calculated through several unit times,
• wherein the detecting an abnormal respiration further includes adding the maximum peak-to-valley value during the current unit time to the moving average of peak-to-valley to recalculate an updated moving average of peak-to-valley values, when the difference between the maximum peak-to-valley value during the current unit time and the previously-calculated moving average of peak-to-valley values does not exceed the first reference value, and
• wherein the detecting an abnormal respiration further includes adding the number of peaks during the current unit time to the moving average of the number of peaks to recalculate an updated moving average of the number of peaks, when the difference between the number of peaks during the current unit time and the previously-calculated moving average of the number of peaks during a unit time does not exceed the second reference value.