An Energy-Efficient Healthcare Chest Patch Interface with Multi-Domain On-Sensor Computing and Inter-Sensor Windowing

Recent wearable devices for heterogeneous applications including physiological health status [1]–[5] or harmful environments [5], [6] require appropriate readout circuits for various signals. Their operating time needs to be extended by reducing power consumption or minimizing the amount of data transferred during wireless communication. A proposed chest patch device and its interface IC are configured to allow detection of multi-domain signals in forms of optical, electrical, acoustic, and chemical at single body spot, where cardiovascular diseases such as hypertension and arrythmia are diagnosed based on photoplethysmogram (PPG), electrocardiogram (ECG), bio-impedance (BioZ), phonocardiogram (PCG), and hazardous gas are analyzed. For low-power multimodal sensing operation, an inter-sensor window scheme is proposed to reduce the LED power of the PPG channel which is dominant energy-consuming. For minimal computation and communication power, a proposed multi-domain on-sensor edge computing is included to perform three processing modes for the multi-domain signals with their respective preprocessing. A chest-wearable prototype with an in-house adhesive meta patch is manufactured and experimentally verified to provide on-sensor computing functions such as personalized hypertension diagnosis, three-class arrythmia classification, and mixture gas detection.