A 2.1 μW event-driven wake-up circuit based on a level-crossing ADC for pattern recognition in healthcare

Abstract

We report an asynchronous wake-up circuit with pattern recognition capabilities to duty-cycle biomedical systemon- chips in internet-of-things healthcare applications. The wakeup circuit is based on a level-crossing analog-to-digital converter to reduce both the number of samples and processing power. On-chip voltage references lower the overall system power consumption and reduce the number of external components. The circuit is highly configurable and can be used for a wide variety of applications. A 130nm CMOS prototype is demonstrated with pathological electrocardiogram with premature ventricular contractions, reaching 74:8% of accuracy with only 2:1 W at 1V supply.