An ECG Dataset Representing Real-World Signal Characteristics for Wearable Computers

Abstract — We present an ECG dataset collected in real-world scenarios for wearable devices that includes over 260 recordings of 90-210 seconds that provide guidance for designers to evaluate signal acquisition circuit and system solutions. Several variations on the signal acquisition path are demonstrated, including various sources of interference (baseline wander, motion artifacts, and power line interference), signal path variations (electrode type, coupling method, and common-mode rejection method), and electrode placements (wrist and chest). Based on detailed analysis of signal characteristics under different scenarios, analog front-end (AFE) design recommendations are proposed