A real-time cardiac arrhythmia classification system with wearable electrocardiogram

Long term continuous monitoring of electrocardiogram (ECG) in a free living environment provides valuable information for the prevention on the heart attack and other high risk diseases. Most of the existing devices provide ECG recording in a hospital setting or off-line ECG diagnosis. The design of a real-time wearable ECG monitoring device with cardiac arrhythmia classification system is discussed in this paper. In this system, the wearable sensor node monitors the patient's ECG and motion signal in an unobstructive way that the patient's daily life will not be affected. ECG analog front-end and on-node processing are designed to remove most of the noise and bias, which guarantees an clean and reliable ECG waveform. The ECG waveform is digitalized by an analog-to-digital convertor and transmitted to a smart phone via bluetooth. On the smartphone, the ECG waveform is visualized and a novel layered hidden Markov model is implemented to classify multiple cardiac arrhythmias in real time. This paper evaluates the performance of the hardware design and the classification algorithm.