Research On Daily Blood Oxygen And Heart Rate Detection Based On Signal Quality Evaluation

Cardiovascular disease is one of the main hidden dangers of human health.It has the characteristics of high mortality and high morbidity.Blood oxygen saturation and heart rate are important parameters for identifying cardiovascular diseases.The harm of cardiovascular diseases are reduced by effective daily detection.Photoplethysmography is the main method to detect blood oxygen and heart rate in daily life.However,pulse signal detection is susceptible to noise interference,resulting in poor signal quality,which in turn causes false alarms of monitoring equipment.Therefore,it is of great significance to study the accurate detection of blood oxygen and heart rate in the daily state,so as to realize the monitoring and early warning of the health status of cardiovascular disease and its potential patient population.The main contents of this thesis are as follows:1)Aiming at the problem that finger and wrist commercial oximeters are greatly affected by exercise interference and blood perfusion in daily testing,the blood oxygen and heart rate detection system is designed based on the part behind the ear.The system that uses STM32F103RCT6 as the main control chip,realizes signal collection through MAX30100 sensor,and transmits the data to the upper computer wirelessly.The design of the host computer includes baud rate and port settings,data acquisition,real-time waveform display and data storage.Through the error analysis of the system,it can be known that the errors of blood oxygen and heart rate conform to medical standards,thus ensuring the effectiveness and stability of daily signal acquisition.2)Aiming at the problem of poor pulse signal quality caused by noise interference in the daily state,a method for evaluating the quality of daily pulse signal based on support vector machine and feature fusion is proposed.Firstly,based on the analysis of daily interference,6 sets of static and motion experiments are designed for simulating the daily state of the real human body.After the data were obtained,the experts marked the signal quality level as "good,medium,and poor".Secondly,the seven features that reflect the interference information are extracted.The grid search method is used to realize the optimization of the support vector machine classification model parameters,and use the features and their combinations as input parameters to realize the signal quality classification.Finally,the best feature vector for distinguishing signal quality is obtained by comparing the accuracy of classification.3)Aiming at the problem of low accuracy of blood oxygen and heart rate detection due to the influence of signal quality in daily behavior environment,a blood oxygen and heart rate detection method based on signal quality evaluation is proposed.The trained classification model is used to evaluate the signal quality,the signals with "bad" quality levels are eliminated according to the classification results,and the influence of motion artifacts on the quality "medium" signals is suppressed based on the generalized combined morphology method.Then the peaks and troughs of the quality "good" and "medium" signals are extracted,and blood oxygen and heart rate are obtained.The results of the study show that,compared to directly calculating blood oxygen and heart rate,the error of blood oxygen and heart rate methods based on signal quality evaluation in the resting state is reduced by 1.1% and 0.8 BPM,respectively,and the accuracy rate in the exercise state is improved from 69.5% to 88.4%.Therefore,the proposed method has good anti-interference ability,and provides a new way and idea for the accurate detection of blood oxygen and heart rate in daily unsupervised state.