Digitally assisted analog front-end power management strategy via dynamic reconfigurability for robust heart rate monitoring

This paper presents a reconfiguration methodology to digitally assist a reconfigurable analog front-end (AFE), with the objective of reducing the power consumption of an ECG-based cardiac activity monitoring system, while maintaining an acceptable performance for the desired signal processing. In this study, we focus on the performance of ECG-based heart rate estimation as an example to demonstrate our proposed strategy. Utilizing the consistency and quasi-periodicity of the ECG waveform, two regions are pre-defined based on the prediction of the R peak by a normalized least mean square (NLMS) adaptive filter. The power consumption and performance of the AFE is dynamically reconfigured accordingly. Experimental evaluations show the system can measure heart rate variability (HRV) with an error of 0.5-4 beats/min with the sampling rate reduced from 488 sps to 100 sps and 40 sps for the two regions respectively, bit resolution reduced from 10-bit to 6-bit and noise tolerance substantially relaxed, offering an estimated 62% total power saving.