Research Of Application Of Heart Rate Estimate Based On IMU On Bed

Cardiovascular disease was an important cause of death,some previous research indicated that HR(heart rate)is a strong predictor of cardiovascular disease.The built-in IMU(inertial measurement unit)of the mobile phone has the advantages of high flexibility and popularity,which is conducive to SCG(seismocardiography)acquisition in the daily environment,thereby acheiving the heart rate self-test.It has a high clinical value.Based on the analysis of the shortcomings of some previous SCG-based heart rate acquisition system,this paper proposed an adaptive real-time heart rate extraction algorithm,and an online distributed real-time heart rate computing system based on the IMU built-in mobile phone.This algorithm used a finite-length signal segment as a training sample,and the characteristic waveform model of the heart beat position was extracted by the K-means clustering algorithms,whose optimal k was determined by the CH(Calinski-Harabasz)index.The heartbeat points in the subsequent signals were marked in real-time by correlation coefficient and feature distance.Then a model iteration processing optimized by the random disturbance and the acceptance probability of the simulated annealing algorithm was used to improve the accuracy of the characteristic waveform model and accelerate the convergence of the model iteration process.The algorithm was validated by the CEBS(combined measurement of ECG,breathing and seismocardiogram)database of Massachusetts Institute of Technology.The results showed that,in our study the adaptive real-time heart rate extraction algorithm had a good adaptive and real-time ability.In addition,this algorithm showed a high accuracy.This software developed by Object-C used iPhone as an acquisition terminal,and run on iOS 9 or later operating systems.In the cluster,the Docker technology was used to simplify the deployment of the algorithm,and the Kubernentes technology was used to achieve the cross-node management of the Docker container.At the same time,based on the Etcd technology,the system contains a registration center function and a monitoring center function,and based on the WebHook technology,an automatically deployment system was achieved.The results of stress test on the system in our study showed that the system had a strong feasibility,a dynamic expansion ability,and a high concurrent performance ability.This paper proposed an improved scheme for SCG-based heart rate acquisition system from two aspects of algorithm and system.The experimental verification showed that the scheme has strong feasibility,and achieved higher accuracy and concurrency performance.