Improving Heart Rate Monitoring During Sleep Through Trunk Muscle Artifact Separation from ECG Signals

Abstract

Monitoring heart rate (HR) during sleep is a crucial method for detecting early signs of cardiac and sleep disorders. But regular heart rate monitors and accelerometers often have trouble telling the difference between normal changes in heart rate and changes that happen when you move or use your muscles. This paper proposes a method to isolate trunk muscle signals (ECG-TMS) from ECG recordings to address this issue. This method utilises the discrete wavelet transform (DWT) to get rid of artifacts caused by muscles, which makes heart rate readings more accurate. A controlled experiment involving a healthy adult male was executed using Shimmer3 ECG equipment to gather data in simulated sleep environments. The method of ECG-TMS did a great job of getting clean heart signals, which made the signal-to-noise ratio (SNR) much better and made it possible to reliably find the heart rate even when the body moved a little. This method was better at getting rid of noise and picking up small muscle movements than systems that used accelerometers. These findings suggest that ECG-TMS may be particularly beneficial in sleep research, cardiovascular diagnostics, and wearable monitoring technologies. Further research will aim to validate the system in real-world sleep scenarios and improve it for continuous monitoring in clinical and home settings.