A Novel Adaptive Spectrum Noise Cancellation Approach for Enhancing Heartbeat Rate Monitoring in a Wearable Device

This paper presents a novel approach, adaptive spectrum noise cancellation (ASNC), for motion artifacts removal in photoplethysmography (PPG) signals measured by an optical biosensor to obtain clean PPG waveforms for heartbeat rate calculation. One challenge faced by this optical sensing method is the inevitable noise induced by movement when the user is in motion, especially when the motion frequency is very close to the target heartbeat rate. The proposed ASNC utilizes the onboard accelerometer and gyroscope sensors to detect and remove the artifacts adaptively, thus obtaining accurate heartbeat rate measurement while in motion. The ASNC algorithm makes use of a commonly accepted spectrum analysis approaches in medical digital signal processing, discrete cosine transform, to carry out frequency domain analysis. Results obtained by the proposed ASNC have been compared with the classic algorithms, the adaptive threshold peak detection and adaptive noise cancellation. The mean (standard deviation) absolute error and mean relative error of heartbeat rate calculated by ASNC is 0.33 (0.57) beats.min(-1) and 0.65%, by adaptive threshold peak detection algorithm is 2.29 (2.21) beats.min(-1) and 8.38%, by adaptive noise cancellation algorithm is 1.70 (1.50) beats.min(-1) and 2.02%. While all algorithms performed well with both simulated PPG data and clean PPG data collected from our verity device in situations free of motion artifacts, ASNC provided better accuracy when motion artifacts increase, especially when motion frequency is very close to the heartbeat rate.