Energy-Efficient ECG Event Signal Processing Using Primitive-Based QRS Complex Detection

The low-power QRS complex detection in the Internet-of-Things (IoT) edge node is helpful in decreasing the number of wireless communications required for serverside DSP computation in wearable applications. This paper proposes an energy-efficient ECG signal processing technique using primitive-based QRS complex detection and demonstrates successful realization in commercial microcontrollers with lowpower consumption, small code size and low profile of required data memory space. The on-chip digital signal processing (DSP) algorithm efficiently extracts the complex’s onset and offset based on the QRS complex’s morphological characteristics as important information for medical treatment. The events of these timed features are only transferred to the server computing platform. The proposed method is validated on an Arduino mega microcontroller (MCU) for physical human body signals and shows a 2.75% reduction in power consumption, compared to the server-centric DSP for the sampled raw data.