A dynamically reconfigurable ECG analog front-end with a 2.5 × data-dependent power reduction

This work addresses a fundamental drawback of conventional bio-signal acquisition interfaces by incorporating a reconfigurable analog front-end (AFE) that can be configured to always operate in an optimal power mode through digital assistance. Implemented in 65 nm CMOS, the AFE has tunable noise performance exhibiting an input-referred noise ranging from 1.90-2.91 while consuming 307-769 nW power from a 0.8 V supply. The low activity and quasi-periodicity of bio-signals are exploited using an LMS-based adaptive linear predictive filter for power reduction. An agile, on-the-fly dynamic noise-power reconfiguration strategy is used to achieve data-dependent power savings of 2.5× without compromising the system’s anomaly detection capability or ability to extract pristine ECG features.