Heart Rate Estimation Algorithm Based on Signal Quality Estimation and Kalman Filter
LI Qiao1, Roger G Mark2, Gari D Clifford2, YU Meng-sun3 (1.Institute of Biomedical Engineering, School of Medicine, School of Control Science and Engineering, School of Medicine, Shandong University, Ji'nan Shandong 250012, China; 2.Harvard-MIT Division of Health Sciences & Technology, MIT, Cambridge, MA, USA; 3.Institute of Aviation Medicine, China Air Force, Beijing 100036,China)
Objective: To develop a robust heart rate (HR) estimation method based on the HR estimates derived from multiple electrocardiogram (ECG) leads from intensive care patients. Methods: Heart rate was obtained by an open source QRS detection algorithm. Physiological signal quality indices (SQI) were obtained by analyzing the morphological and statistical characteristics of each waveform and their relationships. Robust HR estimation was obtained by a Kalman filter based upon the SQI adjustment. The state of Kalman filter could be adjusted by the innovation of Kalman filter and the SQI of ECG data. This method was evaluated using more than 6000 hours of simultaneously acquired ECG from a 437 patient subset of the Multi-Parameter Intelligent Monitoring for Intensive Care II database and adding real ECG noise. Results: Compared with other heart rate estimation methods such as the direct estimation or sample and hold method, this method has the smallest root mean square error (rMSE) of heart rate estimation even when high levels of persistent noise occur. Conclusion: This algorithm provides an accurate HR estimation even in the presence of high levels of persistent noise and artifact.