The Algorithms Research Of Hypoperfusion For Pulse Oximeter

The algorithm of hypoperfusion for pulse oximeter is discussed in this thesis. Oxygensaturation is an important parameter of respiratory system and reflects the blood oxygencontent.Pulse oximeter is a usefull clinical mean to monitor SpO2、perfusion index(PI) andpulserate(PR)for its noninvasive、continuous and real-time. When the PI of users isextremely low,the pulsewaves detected will be so weak that useful signal is almost drawnby noise.It’s difficult to calculate the parameters.We designed a pulse oximeter based on STM32and PC.And then presented two newSpO2algorithms for hypoperfusion.The first is based on FFT and autocorrelation. Usingautocorrelation inhibit the noise and obtain the fundamental frequency of the originalsignal through the autocorrelation signal.According to the amplitude frequencycharacteristics of pulsewave, take the fundamental and second harmonic, three harmoniccomponent of original signal to rebuild signals which used to calculate the parameters.Thesecond is autocorrelation modeling method.According to the analytical relationshipbetween autocorrelation signal and original signal,directly calculate the parametersthrough the high SNR autocorrelation signal.We verified the feasibility of the above two algorithms through theory and simulationresults.Also we tested the accuracy and stability of above two algorithms in the conditionof low perfusion. The experimental results show that the two algorithms could calculateSpO2、PI、PR exactly on the condition that the perfusion is above0.2%.At last the secondalgrithom was integrated into our prototype system that improved the accuracy ofparameter calculation greatly to the pulse oximeter.