Biography
Enrollment Date: 2012
Graduation Date：2015
Degree：M.S.
Defense Date：2015.06.04
Advisors：Zhihua Wang
Department：Institute of Microelectronics，Tsinghua University
Title of Dissertation/Thesis：Research on Accelerometer Aided Dynamic ECG Noise Cancellation Method
Abstract:
Dynamic ECG monitor, Holter, can continuously record long term cardiac activity, which plays an important role in non-invasive diagnosis of heart diseases. Compared with normal static ECG monitoring, the dynamic ECG monitor can improve the diagnostic rate of some cardiac diseases such as intermittent cardiac arrhythmias. Nowadays, the wearable characteristic of dynamic ECG monitor has also made it to be used in individual daily healthcare monitoring system. However, dynamic ECG signal is prone to be interfered by electromagnetic interference from the power lines, movement and respiration of wearer, which can generate motion artifact, baseline wandering, power line interference, etc. These interferences would seriously influence the quality of ECG signal, and bring misdiagnosis or miss-diagnosis. Furthermore, during the conventional dynamic ECG monitoring, the patient has to keep a diary of his/her activities, because the doctor need match the information in the diary to diagnose. Keeping a diary during the monitoring makes the monitor inconvenient to use.
In this work, the problems of the existing dynamic ECG monitor have been investigated, and an accelerometer aided wearable dynamic ECG monitoring system is proposed. This system can collect Lead V2 ECG signal and abdominal acceleration. This work is focused upon accelerometer aided dynamic ECG processing algorithms, recognizing different types of daily activity of the patients and enhancing the quality of ECG signal accordingly, employing the data collected by the accelerometer. In order to improve the quality of the dynamic ECG signal, a specified hardware circuit is utilized to remove the power line interference, which can improve the signal quality and do not need to occupy the computing unit; a baseline extraction method based on the gradient weighting function is proposed to cancel the baseline wandering; the motion artifact is removed according to the recognized activity type, using QRS feature points detection and RLS adaptive filter algorithm.
The proposed dynamic ECG monitoring system is composed of a wearable ECG monitor and a device for data saving and processing. The ECG monitor consist an ultra-low power micro controller unit as a controller, a 24-bit low-power ADC for ECG collecting and a high-accuracy and low-power MEMS accelerometer for abdominal acceleration collecting. In this work, a Personal Computer is utilized as the data storage and processing device, and the wearable ECG monitor communicates with the PC through a low power Bluetooth Module. The monitor is low-power, compacted, small size and light weight, the built-in battery can support about 48 hours continuously dynamic ECG monitoring.
In order to verify the system’s usability and robustness, data from a standard ECG simulator as well as 13 volunteers are collected. The result of the standard ECG simulator shows that the proposed ECG monitoring system can collect the normal ECG signal and record all the ECG characteristics. The result of volunteers shows that this system can recognize daily activities such as coughing, squatting, sitting, walking, etc. And the system can also improve the ECG signal’s R-peak ratio by 7.1 times, and improve the average SNR (signal to Noise Ratio) of the ECG signal to 45.93dB.