A Machine Learning Based Heart Rate Monitoring System

With the rapid development of economy and the accelerated pace of life,human health is threatened by various diseases,and heart diseases are the number one killer of human health.Nowadays,the incidence rate of heart diseases,especially cardiovascular diseases,has increased year by year,and patients tend to be younger.The frequent pattern of hospital visits is no longer applicable to the present.Therefore,medical devices need to move towards convenience.For patients with heart diseases,there are two types of heart rate monitors in the market at present: one is a fixed bedside monitor,which is mainly used in hospitals and other places,with stable performance,including real-time display of ECG,intelligent analysis of ECG data and other functions;the other is a portable heart rate monitor with embedded microcontroller as the core,which is easy to operate But its function is far less powerful than the former.The emergence of artificial intelligence has greatly improved the development of medical industry.In the past,many diseases can only be confirmed by precise analysis with the aid of instrument detection.Through big data analysis of artificial intelligence,results can be obtained more quickly and accurately.Using artificial intelligence can effectively improve the accuracy of the latter.Based on the above background,this paper proposes a design scheme of heart rate monitoring system based on machine learning.The system can collect ECG data,analyze the big data of the data results,and effectively monitor the abnormal condition of the heart.This paper mainly completed the following work:1.The electrocardiograph acquisition circuit is designed and constructed,which includes lead circuit,signal amplification circuit,filter circuit,MCU,communication circuit,etc.The circuit uses MCU as the control core to complete data acquisition,data storage,data transmission and other functions,so that the upper computer can receive the collected ECG data.At the same time,the power circuit,reset circuit,download and debug circuit and other peripheral circuits ensure the normal collection work.2.Read and process the collected data and other ECG data,including ECG waveform conversion,data feature extraction,data preprocessing,data dimensionality reduction,data set division and so on.3.The ECG anomaly classification model for monitoring heart rate is established.The process includes model selection,model construction,model training,parameter optimization and result ratio equivalence.According to the test results of the front-end data,it is confirmed that the final ECG anomaly classification model is very stable and reliable,which can well complete the function of heart rate monitoring.