STUDY OF WEARABLE ECG/EKG DEVICE
Disciplines
Computer and Systems Architecture | Computer Engineering | Digital Circuits | Digital Communications and Networking | Hardware Systems | Other Computer Engineering
Abstract (300 words maximum)
A surge of electrical signal flows in the body, with every heartbeat. This electrical signal can be captured by an electrode with conducting capacity and processed and finessed to produce a noise-free, rhythmic heartbeat signal known as an electrocardiogram or EKG. This is the topic of study in this directed research, to find ways of collecting the data by a small device and study the accuracy of the data obtained. The next step in this study is to use deep learning and Convolutional Neural Networks to study the heartbeat obtained and identify an irregular heartbeat.
In the first part of the semester, we used a prototype board to get the data from the analog frontend such as AD8232, programmed it in C, and used MATLAB to process and analyze the data. An alternate method to get the signals was with the help of a PC sound card and LM741 op-amp. This set of reading obtained from the sound card was a noisy signal and thus a low-pass filter with a cut-off frequency of 400 Hz was set to get the proper data reading. Data were obtained with the help of generic-grade electrodes attached to the chest and to the abdomen. Since the topic is a wearable device to obtain the reading of heartbeat, the attempt of the sound card was made to make the device as small as possible. In the second part of the semester, we used a characteristic identifying technique by using a filter to determine peaks of P-wave, R-Wave, and T-wave in the data set, followed by providing padding and passing it through the hidden 2 layers of the deep network. The purpose of the study is to see the irregularity in these peaks and thus determine the heart condition. Thus the result can help in detecting the symptoms of Arrhythmia.
Academic department under which the project should be listed
Computer Engineering
Primary Investigator (PI) Name
Jeffrey Yiin
STUDY OF WEARABLE ECG/EKG DEVICE
A surge of electrical signal flows in the body, with every heartbeat. This electrical signal can be captured by an electrode with conducting capacity and processed and finessed to produce a noise-free, rhythmic heartbeat signal known as an electrocardiogram or EKG. This is the topic of study in this directed research, to find ways of collecting the data by a small device and study the accuracy of the data obtained. The next step in this study is to use deep learning and Convolutional Neural Networks to study the heartbeat obtained and identify an irregular heartbeat.
In the first part of the semester, we used a prototype board to get the data from the analog frontend such as AD8232, programmed it in C, and used MATLAB to process and analyze the data. An alternate method to get the signals was with the help of a PC sound card and LM741 op-amp. This set of reading obtained from the sound card was a noisy signal and thus a low-pass filter with a cut-off frequency of 400 Hz was set to get the proper data reading. Data were obtained with the help of generic-grade electrodes attached to the chest and to the abdomen. Since the topic is a wearable device to obtain the reading of heartbeat, the attempt of the sound card was made to make the device as small as possible. In the second part of the semester, we used a characteristic identifying technique by using a filter to determine peaks of P-wave, R-Wave, and T-wave in the data set, followed by providing padding and passing it through the hidden 2 layers of the deep network. The purpose of the study is to see the irregularity in these peaks and thus determine the heart condition. Thus the result can help in detecting the symptoms of Arrhythmia.